A PRACTICAL APPROACH TO NEUROPHYSIOLOGIC INTRAOPERATIVE MONITORING

Husain 00

1/17/08

11:50 AM

Page i

A Practical Approach to Neurophysiologic Intraoperative Monitoring

Husain 00

1/17/08

11:50 AM

Page ii

Husain 00

1/17/08

11:51 AM

Page iii

A Practical Approach to Neurophysiologic Intraoperative Monitoring

Edited by

Aatif M. Husain, MD Department of Medicine (Neurology) Duke University Medical Center Durham, North Carolina

New York

Husain 00

1/22/08

11:17 AM

Page iv

Acquisitions Editor: R. Craig Percy Cover Designer: Aimee Davis Indexer: Joann Woy Compositor: TypeWriting Printer: Edwards Brothers Incorporated Visit our website at www.demosmedpub.com © 2008 Demos Medical Publishing, LLC. All rights reserved. This book is protected by copyright. No part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Library of Congress Cataloging-in-Publication Data A practical approach to neurophysiologic intraoperative monitoring / edited by Aatif M. Husain. p. ; cm. Includes bibliographical references and index. ISBN-13: 978-1-933864-09-9 (pbk. : alk. paper) ISBN-10: 1-933864-09-5 (pbk. : alk. paper) 1. Neurophysiologic monitoring. I. Husain, Aatif M. [DNLM: 1. Monitoring, Intraoperative—methods. 2. Evoked Potentials—physiology. 3. Intraoperative Complications—prevention & control. 4. Trauma, Nervous System—prevention & control. WO 181 P895 2008] RD52.N48P73 2008 617.4'8—dc22 2008000450 Medicine is an ever-changing science undergoing continual development. Research and clinical experience are continually expanding our knowledge, in particular our knowledge of proper treatment and drug therapy. The authors, editors, and publisher have made every effort to ensure that all information in this book is in accordance with the state of knowledge at the time of production of the book. Nevertheless, this does not imply or express any guarantee or responsibility on the part of the authors, editors, or publisher with respect to any dosage instructions and forms of application stated in the book. Every reader should examine carefully the package inserts accompanying each drug and check with a his physician or specialist whether the dosage schedules mentioned therein or the contraindications stated by the manufacturer differ from the statements made in this book. Such examination is particularly important with drugs that are either rarely used or have been newly released on the market. Every dosage schedule or every form of application used is entirely at the reader’s own risk and responsibility. The editors and publisher welcome any reader to report to the publisher any discrepancies or inaccuracies noticed. Made in the United States of America 08 09 10

5 4 3 2 1

Husain 00

1/17/08

11:51 AM

Page v

This book is dedicated with love and respect to my parents, Mairaj and Suraiya Husain, who have given me the foundations to do what I have done, what I am doing, and what I will ever do.

Husain 00

1/17/08

11:51 AM

Page vi

Husain 00

1/17/08

11:51 AM

Page vii

Contents

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv

I BASIC PRINCIPLES 1. Introduction to the Operating Room . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Kristine H. Ashton, Dharmen Shah, and Aatif M. Husain

2. Basic Neurophysiologic Intraoperative Monitoring Techniques . . . . . . . . . . 21 Robert E. Minahan and Allen S. Mandir

3. Remote Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Ronald G. Emerson

4. Anesthetic Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Michael L. James

5. Billing, Ethical, and Legal Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Marc R. Nuwer

6. A Buyer’s Guide to Monitoring Equipment . . . . . . . . . . . . . . . . . . . . . . . . . 73 Greg Niznik

II CLINICAL METHODS 7. Vertebral Column Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 David B. MacDonald, Mohammad Al-Enazi, and Zayed Al-Zayed

8. Spinal Cord Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Thoru Yamada, Marjorie Tucker, and Aatif M. Husain vii

Husain 00

viii

1/17/08

11:51 AM

•

Co n te n ts

Page viii

9. Lumbosacral Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Neil R. Holland

10. Tethered Cord Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Aatif M. Husain and Kristine H. Ashton

11. Selective Dorsal Rhizotomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 Daniel L. Menkes, Chi-Keung Kong, and D. Benjamin Kabakoff

12. Peripheral Nerve Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Brian A. Crum, Jeffrey A. Strommen, and James A. Abbott

13. Cerebellopotine Angle Surgery: Microvascular Decompression . . . . . . . . . 195 Cormac A. O’ Donovan and Scott Kuhn

14. Cerebellopotine Angle Surgery: Tumor . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 Dileep R. Nair and James R. Brooks

15. Thoracic Aortic Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Aatif M. Husain, Kristine H. Ashton, and G. Chad Hughes

16. Carotid Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 Jehuda P. Sepkuty and Sergio Gutierrez

17. Epilepsy Surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 William O. Tatum, IV, Fernando L. Vale, and Kumar U. Anthony Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303

Husain 00

1/17/08

11:51 AM

Page ix

Foreword

in the operating room, at least during critical periods of surgery when a neural lesion was likely to occur. Our medicolegal experiences have led us to examine many surgical records in which there was an adverse outcome. Too often the operating surgeon was not informed promptly, usually because response amplitude diminished progressively until it was not clear whether a response was present or not. The surgeon should be aware that false positives are common. He or she should be aware that not all neurophysiologic changes are the result of surgical manipulations. We advised technologists that the surgeon should be informed immediately of the difficulty and the possibility or even likelihood that the cause was purely technical. Then as now, rapid troubleshooting is imperative. It is the surgeon, not the technologist, who knows the anatomy of the current surgery. Thus a change in cortical and/or subcortical responses in spinal surgery may be of little concern if the surgeon is taking a bone graft and has not been near the spine for many minutes. The avoidance of some anesthetic agents as well as the study and documentation of the relative resistance of far-field subcortical responses to anesthetic changes, commonly altering near-field cortical responses, greatly enhanced the utility of evoked potentials (EP) in the operating room setting. Early in the application of NIOM, with the exception of direct cord stimulation, the study of EMG (both spontaneous and evoked responses) was in its infancy. Now, motor responses from

N

europhysiologic intraoperative monitoring (NIOM) is undertaken in an effort to reduce neurological morbidity associated with some types of surgeries. Postoperative outcomes of some orthopedic, neurosurgical, and otologic surgeries reveal a significant incidence of adverse outcomes with intraoperatively acquired neurologic lesions. Presumably there is a time in the development of many lesions when the pathologic process leading to a lesion is reversible. Initially it seemed reasonable that although the patient cannot be directly examined during the surgical procedure, that demonstration of changes in electrophysiology might give warnings that pathophysiologic changes were taking place. If the changes were corrected, the surgically induced lesion could be avoided. These assumptions proved to be true, and various animal studies revealed that neural damage was accompanied by changes of evoked response amplitude, latency, or both and/or changes of evoked and spontaneous electromyographic (EMG) activity. Nonpathologic factors, such as temperature and anesthetic type and depth, cause changes in activity, which may be misinterpreted as an impending neural lesion. Unfortunately such changes may also be due to purely technical factors. When changes are seen, the monitoring technologist should begin a search for technical causes and simultaneously inform the monitoring neurophysiologist and surgeon of the changes. Initially, the technology provided only limited capacity for remote monitoring, requiring the neurophysiologist to be present ix

Husain 00

x

1/17/08

•

11:51 AM

Page x

Fo rewo rd

stimulation techniques, including transcranial electrical and magnetic stimulation, are no longer investigative but are routine in many centers. The technical aspects of electroencephalography (EEG), EP, and NIOM are now largely resolved. Guidelines were published by the American EEG Society in 1986, 1988, 1994, and 2006, which represented consensus knowledge and approaches to these techniques. In the mid-1990s we were introduced to a new postdoctoral fellow in clinical neurophysiology. Aatif Husain was one of the first to take a combined EEG/Epilepsy/EMG fellowship at Duke University Medical Center. To us he expressed a keen interest in NIOM, and he was one of the few fellows to respond to our invitations to come to the operating room for an orientation to NIOM. He would arrive in the operating room at 6:00 A.M. to watch the technical staff prepare patients for monitoring. In a short time he was asking probing questions and assisting in “setting up” the patients for surgery. With his training in EMG, it was not long before Aatif was making suggestions as how to better monitor EMG activity when the surgeon deemed such monitoring to be appropriate. Those who have experienced the operating room environment know well the potential pitfalls related to interfacing with operating room staff. When the monitoring team sets up a new monitoring service and intrudes with additional personnel and equipment, they are not always welcomed with open arms. As in the early days of EEG, technologists and interpreters of EP and NIOM were either selftaught or the product of “on-the-job training” in centers where such diagnostic procedures were done. We were training residents, fellows, and technologists informally, usually structured by the number of trainees present. In a short time we found requests for technologist and interpreter training to be almost overwhelming. Space and the time demands of Duke University Medical Center

allowed us to offer only extended weekend courses. We realized the inadequacy of such training, but too little training was considered preferable to no training. Interpreter classes on EP in general and NIOM in particular could be larger than technologist classes, in part because lab time was not offered. We felt that technologists required a lab exposure, and thus their classes were limited to 20 attendees. Dr. Husain attended many of these technician training sessions, realizing that technologist training was both important and necessary. Subsequently, Dr. Husain championed the long Duke tradition of emphasizing intensive training for technologists, and he has actively supported the American Society of Electroneurodiagnostic Technologists (ASET) courses and the American Board of Registration in Electroencephalographic and Evoked Potential Technology (ABRET) by serving initially as an invited examiner and later as an elected board member. Course offerings in these topics were and are now available at numerous medical centers and national technologist and physician societies have incorporated EP and NIOM into their course offerings at their annual meetings. For physicians, the American Board of Clinical Neurophysiology (ABCN) developed a special NIOM certification examination and the American Board of Psychiatry and Neurology (ABPN) developed the Subspecialty Examination in Clinical Neurophysiology, which incorporated NIOM questions. Early on there was a continuing need to educate surgeons and anesthesiologists on strategies for enhancing NIOM techniques, because cooperation between the surgical and monitoring teams is essential. Dr. Husain’s text will no doubt serve well to promote such cooperation. On the technologist front, ABRET offered the first examination of Certification in Neurophysiologic Intraoperative Monitoring (CNIM) in 1996. Since that time, Aatif’s career in clinical neurophysiology has progressed impressively; his publications alone attest to his broad

Husain 00

1/17/08

11:51 AM

Page xi

Foreword

interest in the field. Aatif directs Duke’s NIOM program. We have retired from the active practice of medicine and left clinical neurophysiology and NIOM to younger eyes and minds. On all these levels, from local teaching to national society offerings to national board examinations, Dr. Husain has contributed his time and expertise to the benefit of trainees at all experience levels. We feel that we must all be trainees; otherwise we close our minds to new information. We share Dr. Husain’s enthusiasm for the field and this text. No text can solve all of the training issues of such a broad audience, but this work is clearly an important step forward. We are honored to

•

xi

have been asked to write this foreword and to have participated in the formative period of this fascinating field. C. William Erwin, MD Professor Emeritus Duke University Medical Center Past President ABRET and ACNS Past Chairman, ABCN Andrea C. Erwin, R EP T, CNIM Past Manager Clinical Neurophysiology Laboratories Duke University Medical Center Past ABRET Section Chair EP Examining Committee

Husain 00

1/17/08

11:51 AM

Page xii

Husain 00

1/17/08

11:51 AM

Page xiii

Preface

there remains a shortage of qualified neurophysiologists and technologists to meet the demand. Education in NIOM is seldom available in traditional residency, fellowship, or other training programs. Technologists also learn on the job and often have no formal curriculum to follow. Textbooks on the subject are few and often limited in scope to neurosurgical and orthopedic procedures. Learning about medicolegal issues surrounding NIOM is even more difficult. With this book I hope to bridge some of the gaps on NIOM education and clinical practice. It is divided in two main sections. In the first section, “Basic Principles,” the reader will learn the basics of NIOM. This includes not only the modalities used in monitoring, but also topics not usually discussed. The novice will find an introduction to the operating room, where basics about sterile technique and the various machines that might be encountered are discussed. Chapters on remote monitoring, billing, and ethical issues will help those trying to set up laboratories. All will benefit from the buyer’s guide to NIOM machines. The second part of the book, “Clinical Methods,” reviews the use of NIOM in various types of surgeries. The chapters present the basics of anatomy, physiology, and surgery of the various procedures so that the reader will better understand what happens during these procedures. Details of the monitoring modalities and their interpretive criteria are then presented. Whenever possible, data supporting the use of NIOM in the particular type of surgery is also presented. A unique feature of every chapter is

S

eldom has a field in medicine grown as rapidly as neurophysiologic intraoperative monitoring (NIOM). Twenty-five years ago, few people had heard of NIOM. About 20 years ago, the initial seminal papers appeared, documenting the utility of somatosensory and brainstem auditory evoked potential monitoring in reducing the morbidity of scoliosis and brainstem surgery. Fifteen years ago, NIOM was available only in large, academic hospitals, often with equipment cobbled together by experienced neurophysiologists. As recently as 10 years ago, there were few educational offerings where one could learn more about NIOM. The last decade has seen remarkable advances in NIOM. Not only have new modalities, such as motor evoked potentials, become available, but the equipment used for monitoring has improved considerably. Advances in telecommunications have enabled remote real-time review of NIOM data by neurophysiologists. Educational programs and courses in NIOM are much more common and are offered during meetings of many neurophysiology societies. Trained technologists have done much to further advance the field. Research has documented the utility of many types of NIOM in reducing the morbidity not only in orthopedic surgery and neurosurgery but also otolaryngologic, vascular, and cardiothoracic surgery. Surgeons, anesthesiologists, neurologists, and other healthcare professionals now accept the value of NIOM in making many types of surgeries safer. Although most large academic and many community hospitals now offer NIOM services, xiii

Husain 00

x iv

1/17/08

•

11:51 AM

Page xiv

Pre f a c e

that it is written not only by an experienced neurophysiologist but also a senior technologist who has contributed to a technical section. This section reviews practical information about the logistics of monitoring a particular type of surgery, and it will be especially useful for technologists. Many different professionals will find this book useful. Neurophysiology trainees and technologists will benefit in reading it in its entirety. Experienced neurophysiologists will find it useful every day in clinical practice, as information about many different types of surgery and monitoring is easily accessible. Surgeons and anesthesiologists will find useful information that will help them understand what the NIOM team does. Laboratory managers will find material that will help them set up laboratory procedures and protocols. There are many individuals that have contributed to this book and must be recognized. Foremost, I am extremely grateful to my col-

leagues who contributed to this text. They have helped create a truly unique book. R. Craig Percy and his colleagues at Demos Medical Publishing have been incredibly supportive of this project and must be recognized and appreciated for their foresight in seeing the need for such a book. A special note of thanks is due to the many technologists, neurophysiologists, and surgeons at Duke University Medical Center who have taught me extensively. Special mention is due to Dr. C. William Erwin and Andrea Erwin, my mentors, without whom I would not have been introduced to NIOM. Of course, none of this would be possible without our patients, who have done the most to advance this field. Finally, I am indebted to my wife, Sarwat, and children, Aamer and Aayaz, for enduring my absenteeism for the many, many nights and weekends I worked on this book. Aatif M. Husain

Husain 00

1/17/08

11:51 AM

Page xv

Contributors

James A. Abbott, R ED T

Ronald G. Emerson, MD

Department of Neurology Mayo Clinic College of Medicine Rochester, Minnesota Chapter 12: Peripheral Nerve Surgery

Department of Neurology The Neurological Institute Columbia University College of Physicians and Surgeons New York, New York Chapter 3: Remote Monitoring

Kumar U. Anthony, R EEG/EP T, CNIM, R PSG T

Mohammad Al-Enazi

Department of Neurology Tampa General Hospital Tampa, Florida Chapter 17: Epilepsy Surgery

Department of Neurosciences King Faisal Specialist Hospital & Research Centre Riyadh, Saudi Arabia Chapter 7: Verterbral Column Surgery

Kristine H. Ashton, R EEG T, CNIM Department of Medicine (Neurology) Duke University Medical Center Durham, North Carolina Chapter 1: Introduction to the Operating Room Chapter 10: Tethered Cord Surgery Chapter 15: Thoracic Aortic Surgery

Sergio Gutierrez, MD Department of Neurology Johns Hopkins University and Hospital Baltimore, Maryland Chapter 16: Carotid Surgery

Neil R. Holland, MB BS

James R. Brooks, R PSG T, CNIM

Neurology Specialists of Monmouth County West Long Branch, New Jersey Department of Neurology Drexel University College of Medicine Philadelphia, Pennsylvania Chapter 9: Lumbosacral Surgery

Epilepsy Center The Neurological Institute Cleveland Clinic Cleveland, Ohio Chapter 14: Cerebellopontine Angle Surgery: Tumor

G. Chad Hughes, MD

Brian A. Crum, MD

Department of Surgery (Thoracic and Cardiovascular) Duke University Medical Center Durham, North Carolina Chapter 15: Thoracic Aortic Surgery

Department of Neurology Mayo Clinic College of Medicine Rochester, Minnesota Chapter 12: Peripheral Nerve Surgery

xv

Husain 00

x vi

1/17/08

•

11:51 AM

Page xvi

Co n tr i b u to r s

Aatif M. Husain, MD

Allen S. Mandir, MD, PhD

Department of Medicine (Neurology) Duke University Medical Center Neurodiagnostic Center Veterans Affairs Medical Center Durham, North Carolina Chapter 1: Introduction to the Operating Room Chapter 8: Spinal Cord Surgery Chapter 10: Tethered Cord Surgery Chapter 15: Thoracic Aortic Surgery

Department of Neurology Georgetown University Washington, DC Chapter 2: Basic Neurophysiologic Intraoperative Monitoring Techniques

Michael L. James, MD Departments of Anesthesiology and Medicine (Neurology) Duke University Medical Center Durham, North Carolina Chapter 4: Anesthetic Considerations

D. Benjamin Kabakoff, R EEG T, CNIM, R PSG T Department of Clinical Neurophysiology Baptist Memorial Hospital Memphis, Tennessee Chapter 11: Selective Dorsal Rhizotomy

Chi-Keung Kong, MBChB, FHKAM (Paed) Department of Paediatrics The Chinese University of Hong Kong, Hong Kong, People’s Republic of China Chapter 11: Selective Dorsal Rhizotomy

Scott Kuhn, BS, CNIM Department of Diagnostic Neurology Wake Forest University Baptist Medical Center Winston-Salem, North Carolina Chapter 13: Cerebellopontine Angle Surgery: Microvascular Decompression

David B. MacDonald, MD, FRCP(C), ABCN Department of Neurosciences King Faisal Specialist Hospital & Research Centre Riyadh, Saudi Arabia Chapter 7: Verterbral Column Surgery

Daniel L. Menkes, MD Department of Neurology University of Tennessee Health Sciences Center at Memphis Department of Clinical Neurophysiology Baptist Memorial Hospital Memphis, Tennessee Chapter 11: Selective Dorsal Rhizotomy

Robert E. Minahan, MD Department of Neurology Georgetown University Washington, DC Chapter 2: Basic Neurophysiologic Intraoperative Monitoring Techniques

Dileep R. Nair, MD Epilepsy Center The Neurological Institute Cleveland Clinic Cleveland, Ohio Chapter 14: Cerebellopontine Angle Surgery: Tumor

Greg Niznik, MSc, CNIM, DABNM Allied Neurodiagnostics Atlanta, Georgia Chapter 6: A Buyer’s Guide to Monitoring Equipment

Marc R. Nuwer, MD, PhD Department of Neurology UCLA School of Medicine Department of Clinical Neurophysiology UCLA Medical Center Los Angeles, California Chapter 5: Billing, Ethical, and Legal Issues

Cormac A. O’Donovan, MD Department of Neurology Wake Forest University Baptist Medical Center Winston-Salem, North Carolina Chapter 13: Cerebellopontine Angle Surgery: Microvascular Decompression

Husain 00

1/17/08

11:51 AM

Page xvii

Contributors

Jehuda P. Sepkuty, MD Epilepsy Center/Clinical Neurophysiology Swedish Neuroscience Institute Seattle, Washington Johns Hopkins University and Hospital Baltimore, Maryland Chapter 16: Carotid Surgery

Dharmen Shah, MD Department of Medicine (Neurology) Duke University Medical Center Durham, North Carolina Chapter 1: Introduction to the Operating Room

Jeffrey A. Strommen, MD Department of Physical Medicine and Rehabilitation Mayo Clinic College of Medicine Rochester, Minnesota Chapter 12: Peripheral Nerve Surgery

William O.Tatum, IV, DO Department of Neurology Tampa General Hospital and University of South Florida Tampa, Florida Chapter 17: Epilepsy Surgery

•

x vii

Marjorie Tucker, CNIM, R EEG/EP T, R NCS T Department of Neurology Roy J. and Lucille A. Carver College of Medicine University of Iowa Hospital and Clinics Iowa City, Iowa Chapter 8: Spinal Cord Surgery

Fernando L.Vale, MD Department of Neurosurgery Tampa General Hospital and University of South Florida Tampa, Florida Chapter 17: Epilepsy Surgery

Thoru Yamada, MD Department of Neurology Roy J. and Lucille A. Carver College of Medicine University of Iowa Hospital and Clinics Iowa City, Iowa Chapter 8: Spinal Cord Surgery

Zayed Al-Zayed, MD, FRCS Department of Orthopedics King Faisal Specialist Hospital & Research Centre Riyadh, Saudi Arabia Chapter 7: Verterbral Column Surgery

Husain 00

1/17/08

11:51 AM

Page xviii

Husain 00

1/17/08

11:51 AM

Page xix

A Practical Approach to Neurophysiologic Intraoperative Monitoring

Husain 00

1/17/08

11:51 AM

Page xx

Husain 01

1/17/08

11:51 AM

I

Page 1

Basic Principles

Husain 01

1/17/08

11:51 AM

Page 2

Husain 01

1/17/08

11:51 AM

Page 3

1

Introduction to the Operating Room

Kristine H. Ashton Dharmen Shah Aatif M. Husain cussed first, followed by a review of how surgeons and other staff should be attired.

T

he operating room can be an intimidating environment for most nonsurgical healthcare personnel. Lack of operating room knowledge, unfamiliarity with aseptic techniques and sterile field, and minimal surgical training contribute to the anxiety. This basic knowledge is critical for neurophysiologic intraoperative monitoring (NIOM) staff, since it helps them to understand the roles of various personnel and how their jobs dovetail with those of others. Familiarity with the personnel, instrumentation, and environment of the operating room can help make NIOM successful. This chapter serves as an introduction to the operating room, discussing aseptic technique, personnel, equipment likely to be encountered, and the basics of preparation, safety, and cleanup related to NIOM.

Sterilization Methods Sterilization is a process by which objects are cleansed of living organisms, including spores. It is important to differentiate sterilization from disinfection. Disinfection reduces the number of viable organisms, whereas sterilization kills all organisms. Common disinfectants include alcohol and phenolic compounds. However, sterilization—not disinfection—is necessary in the operating room. Three methods of sterilization are available: thermal, chemical, and radiation. Each method uses different devices to measure and ensure the integrity of the sterilization. Knowing about methods of sterilization is important for NIOM staff, as instruments needing to be sterilized must be packed according to the method that will be used to sterilize them.

ASEPTIC TECHNIQUE “Aseptic technique” refers to practices used to minimize the patient’s exposure of pathogens, usually one undergoing a surgical procedure. This technique involves not only cleaning surgical instruments so that they are free of pathogens but also routines followed by surgeons and operating room personnel to minimize the spread of microorganisms. In this section, sterilization techniques are dis-

Thermal Sterilization Heat is a reliable method of sterilization. Moisture, in the form of steam, combined with heat hastens the process. This process causes denaturation and coagulation of enzymes and protein systems of microorganisms. All living organisms are killed with 3

Husain 01

4

1/17/08

•

11:51 AM

Page 4

SE C T I ON I : B a s i c P r i n c i ple s

steam at a temperature of 115°C (240°F) for 15 minutes (1). Higher temperatures require less time. The exact temperature and time needed to thermally sterilize an object depends on its size and what else is placed in the sterilization chamber. Steam sterilizers are often referred to as autoclaves. Three types of steam pressure sterilizers are commonly used: prevacuum sterilizers, gravity displacement sterilizers, and flash/high-speed pressure sterilizers. As name indicates, the prevacuum sterilizer vacuums air for several minutes prior to entry of the steam into the chamber. Because of the vacuum created, as soon as steam enters the chamber it penetrates to the center of packages placed in the chamber. A temperature between 132° and 141°C (270° and 285°F) at 27 lb/in2 for 15 to 30 minutes is needed for sterilization in prevacuum sterilizers (1). A gravity displacement sterilizer has two openings. Steam enters the chamber through one opening, displacing heavy air out of the chamber from the other opening. Once the heavy air is removed from the chamber, the second opening is closed. The steam penetrating objects within the chamber kills the organisms. Gravity displacement sterilizers require a temperature of 121° to 123°C (250° to 254°F) at 15 lb/in2 for 15 minutes to complete sterilization (1). A flash/high-speed pressure sterilizer uses either a prevacuum or gravity displacement cycle. If a prevacuum cycle is used, the temperature requirement is 132° to 134°C (270° to 275°F) at 27 lb/in2 for 3 minutes to kill microorganisms on a nonporous object. For porous objects, the time must be extended to 4 minutes. If the gravity displacement cycle is used, the time should be extended to at least 10 minutes (1). Flash/high-speed sterilization should be used only in emergencies and not to compensate for a shortage of supplies. Electrodes, stimulators, and other objects used in NIOM that require sterilization should be thermally sterilized. Flash/highspeed pressure sterilization should be used

only in emergencies, as when a one-of-a-kind instrument is inadvertently contaminated, such as by dropping on the floor. The advantages of thermal sterilization are that it is fast, easy, inexpensive, and safe. It disadvantages include the need for cleaning the objects prior to sterilization, ensuring the steam contact with items in the chamber, and different standards regarding exposure time and temperature according to object size.

Chemical Sterilization Multiple different chemicals have been utilized for chemical sterilization; however, only those listed by the U.S. Environmental Protection Agency (EPA) as chemical sterilants should be used as such. This list includes ethylene oxide gas, hydrogen peroxide, formaldehyde, and formaldehyde gas. Most of these chemical agents interfere with the metabolism of the organism, leading to cell death and ensuring sterility. Chemical sterilization is most commonly used for items that can easily erode and have low melting points, such as plastic. It is extensively used by commercial companies to prolong the shelf life of their products. In general, items that can be sterilized with steam should not be sterilized with chemical agents. Electrodes and other objects used by the NIOM team needing sterilization should not be chemically sterilized.

Radiation Sterilization Ionizing radiation is mostly used for commercial sterilization. It kills microorganisms by forcefully dislodging electrons from ions and disrupting their deoxyribonucleic acid (DNA). The free electrons from gamma rays penetrate objects to be sterilized; efficacy depends on the electrons penetrating the entire object. Thus thick and dense objects may not get properly sterilized. However, high heat is not involved, so heat-sensitive objects can be sterilized in this manner. Gamma radiation is very expensive, and exposure of healthcare personnel to the radiation must be

Husain 01

1/17/08

11:51 AM

Page 5

CHA P T ER 1: Introd uction to the O pe rating R oom

closely monitored (2). Radiation sterilization is seldom needed by NIOM personnel.

Operating Room Attire Since human skin is harbors living organisms, operating room attire has been created to provide an effective barrier against the spread of these organisms to the patient. To maintain the integrity of the sterile environment of the operating room, street clothes are not allowed into the restricted area, and operating room attire should not be worn outside the operating room. Different clothing requirements exist for those individuals who are scrubbed for a procedure and for those not scrubbed. Individuals who are not scrubbed for the actual surgery have four required components of their attire: body cover, head cover, shoe covers, and mask. Body covers are available in either two-piece pantsuits or one-piece overalls. If a two-piece suit is worn, the shirt should be tucked inside the pants’ drawstring to avoid contact with the sterile field. The required head and shoe covers are usually made of special paper and should be replaced each time the individual exits and enters the restricted area. Two types of face masks are available, conical and cup-shaped. The application of the mask is also critical for its effective use. The upper strings should always be tied at the back of the head, while the lower strings are tied behind the neck. Multiple different personal protective devices are also available—such as aprons, eyewear, nonsterile glove, etc.—however, their use is not mandatory. NIOM personnel must wear the prescribed attire when they are in the operating room. Personnel who are operating or otherwise scrubbed in for the surgery must undergo three processes to become sterile: scrubbing, gowning, and gloving. Scrubbing the hands and arms with special soaps removes many organisms. Operating rooms may have requirements for how long or how many

•

5

brush strokes are necessary for adequate scrubbing. After drying with a sterile towel, a sterile gown is immediately put on. Details about gowning and gloving are not be discussed further, as NIOM personnel will generally do not need to be scrubbed. Such details can be found elsewhere in operating room texts (1,3).

Sterile Field Knowledge of the sterile field is critical as a small mistake can result in a significant disruption to the operating room environment. The sterile field includes the patient and his or her periphery, personnel wearing sterile attire, furniture covered with sterile drapes, and items within the sterile field. The integrity of the sterile field must be maintained by allowing only sterile items to enter or touch it. If there is a question regarding an item’s sterility, it should be considered nonsterile. Sterile gowns worn by the surgeon and the staff assisting him or her are considered sterile in front of the chest, at and above the level of the sterile field but not below it. Also, the area 2 inches above the elbows to the cuffs is considered sterile. The back of the gown is considered nonsterile, as it is not under close observation. Persons wearing sterile attire should keep their hands at or above waist level or at the level of the sterile field. All surgical team members should maintain the same height in relation to other surgical team member to protect the integrity of the sterile field. A drastic change in the position of one surgical team member, such as the sitting position, will require a change in position of all other members of the surgical team as well (3). Sterile tables are considered sterile only on the top areas. Drapes extending below the operating tables are considered nonsterile. The NIOM personnel must secure all electrical cords with nonperforating clips to prevent them from sliding onto the sterile areas. Sterile items on the table are continuously observed and maintained by the main scrub

Husain 01

6

1/17/08

•

11:51 AM

Page 6

SE C T I ON I : B a s i c P r i n c i ple s

nurse. The circulating nurse (discussed later) is responsible for getting instruments and supplies that are not on the sterile tray of instruments. He or she will then hand the instrument to the scrub nurse or place them in the sterile field, maintaining integrity of the object and the sterile field. If the item is wrapped in a peel package, the edges of the package should be rolled over the presenter’s hands and the inner sterile contents presented to the scrub nurse. This must be done in a manner that avoids contamination of the object, the scrub nurse, and the sterile field. Nonsterile persons should never reach over the sterile field to transfer the sterile items. On the other hand, a sterile person should never reach over a nonsterile area (1,3). The NIOM staff must be able to follow the sterile technique and learn to transfer objects such as electrodes and stimulators in a sterile manner. If they are unable or not sure how to do so, they should consider using the circulating nurse to transfer their supplies to the operating table. Like all other personnel in the operating room, the NIOM staff also have a responsibility of helping keep the sterile field intact and alerting the appropriate personnel (scrub or circulating nurse) if a breach occurs.

PERSONNEL Many different people are likely to be encountered by the NIOM team in the operating room. It is useful to know who these people are, what responsibilities they have, and how they can affect NIOM.

Surgical Team In addition to the attending surgeon, other members of the surgical team may be present in the operating room. Several of these may scrub in the case and assist the surgeon. The NIOM staff must know who is in charge of the case, as this person must be kept

abreast of any changes or other issues with NIOM.

Attending Surgeon The surgical team consists first of an attending surgeon, the physician ultimately responsible for the patient and everything that occurs during the surgery. The attending surgeon decides which patients should undergo NIOM and, in conjunction with the neurophysiologist, decides what type of NIOM to use. Changes in NIOM during surgery must be communicated to the attending surgeon so that appropriate corrective action can be taken if possible. Additionally, any issues with setup of a case should be referred to the attending surgeon so that he or she is aware and can address them if necessary. The attending surgeon is usually not in a position to serve as the interpreter of NIOM data.

Surgical Resident One or more surgical residents (surgeons in training) may be training with the attending surgeon and assist with the case. Senior residents may open and close the case independently. Depending on their skill level, chief residents may perform most of a surgery independently. When the attending surgeon is not in the room, the senior resident is in charge. In such a situation, responsibilities noted above for the attending surgeon are assumed by the resident in charge. Changes in NIOM data must then be communicated to this resident.

Medical Students At some centers, particularly teaching hospitals, medical students may be present and perform certain duties, such as inserting Foley catheters. With the surgeon’s permission, they may scrub in on the case. They generally are not present in the surgery by themselves. Whereas informing medical students about NIOM may be educational for them, it is not a substitute for informing the surgeon in charge. Medical students should

Husain 01

1/17/08

11:51 AM

Page 7

CHA P T ER 1: Introd uction to the O pe rating R oom

not be relied upon to convey NIOM information to the attending surgeon.

Physician Assistants (PAs) and Nurse Practitioners (NPs) PAs and NPs work in large and small hospitals and may scrub in to assist the attending surgeon when necessary. In these situations, PAs and NPs play a role similar to that of a surgical resident. However, they will seldom perform large parts of the surgery independently. If the PA or NP is the only member of the surgical team present in the room, he or she must be kept informed of NIOM changes; if the attending surgeon is present, he or she must be informed.

Anesthesia Team The anesthesia team typically consists of an attending physician and technician who float between several operating rooms and an anesthesia resident or clinical registered nurse anesthetist (CRNA) who is dedicated to one operating room. The monitoring technologist should communicate with the anesthesia team about the best time to apply electrodes and other requirements for NIOM.

Attending Anesthesiologist The attending anesthesiologist is the physician in charge of all aspects of the patient’s anesthesia. Often he or she is responsible for overseeing the anesthetic care of patients in multiple rooms. Although the attending anesthesiologist will not be physically present in the operating room for the duration of the surgery, an anesthesia resident or CRNA will be present at all times. The attending anesthesiologist will usually interview the patient in the preoperative holding area prior to surgery. The NIOM team should communicate any special requirements, such as total intravenous anesthesia (TIVA), with the attending anesthesiologist. When changes in NIOM data occur, in addition to the surgeon being informed, the anesthesiologist

•

7

should also be kept aware in case changes in anesthesia are needed.

Anesthesia Resident The anesthesia resident is an anesthesiologist in training and is responsible for the patient’s immediate care; he or she remains in the operating room with the patient at all times. The monitoring technologist will frequently encounter the resident in the operating room very early in the morning while both are setting up equipment to begin a case. It is courteous to begin communications regarding anesthetic requirements at this time, since the anesthesia team may need to gather different drugs than anticipated. Additionally, they may require additional infusion pumps and some extra time to program these. During the surgery, if NIOM changes are noted, along with the surgeon, the anesthesia resident should be kept apprised of these if the attending anesthesiologist is not in the room.

Clinical Registered Nurse Anesthetist A CRNA is an anesthesia team member who has completed a master’s degree program in nurse anesthesia. This person functions in much the same way as the anesthesia resident and is present with the patient during the entire surgery.

Anesthesia Technician Anesthesia technicians assist the team in setting up the case in a variety of ways. They gather equipment, set up infusion pumps, prepare electrocardiogram leads and blood pressure cuffs, gather intravenous supplies, and do any other job that can help the anesthesia team get a case started. Throughout the case, the technician takes samples to the lab, gets blood, operates machines such as the cell saver, and again carries out other duties as needed.

Nursing Team The nursing team usually consists of two nurses who are assigned to the operating

Husain 01

8

1/17/08

•

11:51 AM

Page 8

SE C T I ON I : B a s i c P r i n c i ple s

room for all of the cases scheduled in that room for the day. They are almost always registered nurses (RNs) and may have a number of other degrees/certifications.

Circulating Nurse The circulating nurse, also known as the circulator, is responsible for keeping track of personnel in the room, retrieving items required during surgery, maintaining documentation, answering the phone, and doing any other task that need to be done by someone who is not sterile. It is important for the monitoring technologist to check in with the circulating nurse at the start of the case. A cooperative relationship about the location of equipment, access to the patient, and draping of electrodes is essential to successful monitoring of a case.

Scrub Nurse The scrub nurse scrubs in prior to the case in order to set up the instruments. He or she remains sterile throughout the case and passes instruments to the surgeons. If the monitoring technologist has any sterile probes, stimulators, electrodes, or cables for the surgeon to use, these are passed in a sterile fashion to the scrub nurse.

Room Attendant The room attendant brings large equipment such as beds or microscopes into the room. He or she also helps to turn and position the patient. The circulating nurse will often call on the room attendant to locate odds and ends, such as extension cords or hair clippers. Attendants are usually assigned to several operating rooms and run between rooms, helping out wherever they can. An experienced room attendant can be an invaluable resource to the monitoring technologist, since he or she often arrives early in the morning to set up rooms, knows how the rooms are arranged and where best to place the NIOM equipment, knows where to find

supplies, and knows a variety of personnel who may be of help.

Allied Health Personnel Depending on the type of surgery, a variety of allied health personnel may be called in to help or may be present in the room throughout the surgery.

Radiology Technician The radiology technician is one of the allied health professionals whom the monitoring tech will encounter often in the operating room. The radiology technician is usually operating one of two types of equipment, either a traditional x-ray or a C-arm machine. It is important for the monitoring technologist to establish communication with the radiology technician about the location of any wires or cords on the patient or bed so that the movement of the radiology machines does not disrupt NIOM.

Perfusionist A perfusionist is present in cardiothoracic surgeries and operates the cardiopulmonary bypass (CPB) machine. The monitoring technologist should be aware that the perfusionist can deliver anesthetic agents through the CPB machine. Consequently, in these cases, not only the anesthesia team but also the perfusionist must be asked what anesthetics the patient will be receiving. The perfusionist is also responsible for inducing hypothermia when needed, and the monitoring technologist must be aware of that as well.

EQUIPMENT A variety of different machines can be found in the operating room. The NIOM team must know not only their function but also the effect they can have on monitoring. Many machines will produce high-amplitude artifacts that must be recognized.

Husain 01

1/17/08

11:51 AM

Page 9

CHA P T ER 1: Introd uction to the O pe rating R oom

•

9

Electrocautery

Cardiopulmonary Bypass Machine

The electrocautery (Figure 1.1), also known simply as cautery, is used in almost every operation. The cautery is used to electrically cut skin and to burn bleeding blood vessels closed. It causes a very large artifact that cannot be averaged through (Figure 1.2). The monitoring technologist will need to “pause” the NIOM machine during active cautery.

The CPB machine is used for aortic surgery (Figure 1.3). In these surgeries, electroencephalography (EEG), somatosensory evoked potentials (SEPs) or motor evoked potentials (MEPs) may be monitored. The perfusionist operates the CPB machine and can administer a variety of anesthetics through it. As noted above, the monitoring technologist must be aware of the drugs being administered through the CPB machine, as they may affect NIOM. The CPB machine can also be used for inducing hypothermia, and when this is being done, the monitoring technologist should make a note, as that will affect the monitoring as well. At times the CPB machine can produce significant artifact that makes interpretation of NIOM difficult. This is particularly true in cases where EEG activity is being recorded at 2 µV/mm to determine electrocerebral inactivity; at such a high sensitivity, artifacts are difficult eliminate.

Microscope and Monitor The microscope is much used in neurosurgery (Figure 1.4). The monitoring technologist must position his or her equipment and cables in such a way as not to interfere with FIGURE 1.1 An electrocautery machine.

FIGURE 1.2 Example of artifact caused by electrocautery. This was seen when tibial somatosensory evoked potentials were being averaged.

FIGURE 1.3 A cardiopulmonary bypass machine. This machine can also be used to administer anesthetic agents and other drugs as well as for inducing hypothermia.

Husain 01

10

1/17/08

•

11:51 AM

Page 10

SE C T I ON I : B a s i c P r i n c iple s

FIGURE 1.4 An operating surgical microscope. During surgery, this microscope will be draped with sterile sheets.

the microscope moving in and out of the field. There is rarely an artifact problem related to the microscope. The microscope will be sterile and draped prior to the start of the case and should not be touched once draped. Often the microscope is connected to a monitor that displays what the surgeon is seeing through the microscope (Figure 1.5). This is helpful, as all operating personnel can then see what is happening in the surgical field.

FIGURE 1.5 A monitor that displays images of the surgical microscope.

Transesophageal Echocardiography During aortic surgery, a procedure called transesophageal echocardiography (TEE) may be performed by the anesthesiologist. During this procedure, the anesthesiologist often has his or her fingers in the patient’s mouth, manipulating the TEE probe. TEE usually does not produce artifact that affects NIOM. However, if MEP monitoring is being performed, it is important to be sure the anesthesiologist does not have his or her fingers in the

Ultrasound Various types of ultrasound are commonly used in the operating room in a variety of surgeries (Figure 1.6). The ultrasound probes are usually operated by the surgeon, but they can be used by the anesthesiologist as well.

Intraoperative Ultrasound A standard ultrasound machine is used to precisely locate tumors or other lesions in the surgical field. The ultrasound probe is covered in a sterile sheath and held by the surgeon. This type of ultrasound is used sporadically during the surgery.

FIGURE 1.6 An ultrasound machine commonly used in the operating room.

Husain 01

1/17/08

11:51 AM

Page 11

CH AP T ER 1: Introd uction to the O pe rating R oom

•

11

patient’s mouth, since a transcranial electrical stimulus will cause the patient to bite down. A surprise MEP stimulation could therefore cause injury to the anesthesiologist. It is best to alert both the surgeon and anesthesiologist when a MEP stimulus is about to be delivered.

Laser A carbon dioxide (CO2) laser is sometimes used during surgeries in which a tumor, such as a lipoma, must be broken down. This is often done in tethered cord surgery. Special CO2 goggles must be worn by operating room personnel to protect their eyes against damage. This laser does not often produce an artifact that affects NIOM.

FIGURE 1.7 The anesthesia work station.

starting points. The x-ray machine does not produce an artifact that interferes with NIOM.

Cavitron Ultrasonic Surgical Aspirator

Fluoroscopy

The cavitron ultrasonic surgical aspirator (CUSA) is another device that uses sound waves to break down tissue, usually tumors. It is used in many different types of surgical procedures. At times it can produce an artifact that makes it difficult to average evoked potentials (EPs).

Fluoroscopy, a type of x-ray system, is used in the operating room; its images can be visualized immediately without having to be processed. It is composed of at least two parts and, if needed, can be used with an injector for visualizing blood vessels. When fluoroscopy is being used, the monitoring technologist should make sure that he or she is appropriately covered with lead protective garments.

Anesthesia Equipment The anesthesia work station includes monitoring equipment, gas delivery devices, a documentation station, storage area, work platform, and a variety of other anesthetic tools (Figure 1.7). This equipment generally does not produce artifact in NIOM. However, the monitoring technologist should learn to identify anesthetics delivered, inspired and expired gases, blood pressures, and temperature, all of which are displayed on the anesthesia equipment.

X-ray A standard x-ray machine is found most often in orthopedic and neurosurgery to locate bone structures so as to determine surgical

C-Arm A C-arm is that part of a fluoroscopy device that is brought in close proximity to the patient and through which the x-rays pass (Figure 1.8). It can be used for diagnostic and interventional purposes. Although a C-arm can be wall-mounted, in the operating room it is often mobile so that it can be positioned appropriately. The C-arm can be angulated in any direction so as to provide the best possible imaging study. Mobile C-arm units provide image recording by either spot-film or digital image acquisition.

Monitor Various image recording devices may be incorporated, including a film changer, cine

Husain 01

12

1/17/08

•

11:51 AM

Page 12

SE C T I ON I : B a s i c P r i n c iple s

FIGURE 1.8 The C-arm. Notice that it is draped with sterile sheets.

FIGURE 1.9 The monitor used to display fluoroscopy images during surgery.

camera, or digital image acquisition for digital subtraction. Most often a monitor is used to display the digital image (Figure 1.9).

The injector is the part of the fluoroscopy system through which a dye can be injected to visualize blood vessels. It is used often in endovascular surgeries. Fluoroscopy seldom causes significant artifact that would prevent adequate NIOM.

ing systems. A magnetic resonance imaging (MRI) study is preloaded in the computer and sensors are attached to the patient. An infrared camera detects the position of the patient’s head. Surgical instruments can also have sensors that make them appear on the MRI, making localization easier. An example of this is the BrainLAB, used in neurosurgical procedures (Figure 1.10). Image-guided surgery systems seldom produce artifact that would hamper NIOM.

Cell Saver

Surgical Table

The cell saver is used to process the patient’s blood and return the red blood cells to the patient (hence the name). At times, especially when grounding is faulty, the cell saver can produce significant artifact that impairs averaging EPs.

Surgical tables are of many different types. They are usually electrically powered and plug into an AC outlet. In addition to moving up and down, they can turn in various directions to help optimize patient positioning. Because they plug into the AC outlet, they can also produce artifact, which can make averaging EPs difficult (Figure 1.11). The artifact can be isolated to the bed if unplugging the bed eliminates the artifact. If this happens, a discussion with the anesthesia team about whether the operating table can remain disconnected must take place.

Injector

Image-guided Surgery Systems Image-guided surgery systems are sophisticated systems that help surgeons identify and locate structures deep to the surface. They work in a manner similar to global position-

Husain 01

1/17/08

11:51 AM

Page 13

CH AP T ER 1: Introd uction to the O pe rating R oom

•

13

FIGURE 1.12 Example of artifact caused by a surgical drill. This was noted during microvascular decompression surgery for trigeminal neuralgia. After exposure, when the drill was no longer in use, the artifact disappeared.

FIGURE 1.10 The BrainLAB, which helps localize structures in the surgical field.

until after the drilling has been completed. This is commonly seen in retromastoid craniotomies, such as those done for cerebellopontine angle surgeries.

PREPARATION

FIGURE 1.11 Example of artifact caused by the operating table. Disconnecting the table eliminated the artifact.

Preparing for a surgery in which monitoring will be used is among the most important things that the NIOM team does. Thorough planning reduces anxiety during the surgery and ensures that the best possible monitoring is performed.

Preoperative Studies Surgical Drill Many surgical procedures involving the skull and other bones require the use of a handheld drill by the surgeon. The appearance of such a drill is similar to that of a drill used around the home; however, they are made to surgical specifications and have special drill bits. When a drill is being used, significant artifact is produced, so that EP monitoring is not possible (Figure 1.12). Many times it is best to stop averaging Es

Preoperative studies are EP or EEG studies performed at least one day prior to surgery. Although it is not always possible to perform preoperative studies, they should be considered if and when possible. The primary reason for performing a preoperative study is verification that the patient has responses that can be monitored during surgery. If any abnormalities exist, these are identified and the surgeon is notified prior to the surgery. This process also allows for a greater degree of confidence in the baseline responses in the operating room.

Husain 01

14

1/17/08

•

11:51 AM

Page 14

SE C T I ON I : B a s i c P r i n c iple s

In certain centers, patients may be screened for preoperative studies based on their medical history. Those with a higher likelihood of abnormal or absent responses may be candidates for preoperative studies. These would include patients with tumors of the central nervous system, patients with peripheral neuropathy, etc.

Contraindications to NIOM There are very few absolute contraindications to NIOM. The presence of a cardiac pacemaker and cranial and other implants is a contraindication to MEP monitoring. In the past, a history of epilepsy was also considered a contraindication to MEP monitoring. However, recent data suggest that the risk of provoking seizures with MEP is very low (4). There are no contraindications to other types of monitoring.

Supplies Some of the supplies the monitoring technologist needs can be found in the operating room. Most of them, however, will need to be

brought in by the technologist. Usually the technologist will carry a small toolbox, apron, or bucket with supplies needed for any situation. The most important of all is, of course, electrodes. Available electrode types include disposable needles, disposable foam pads, reusable gold cups, corkscrew needles, and a variety of other electrodes found in many supply catalogs. A list of supplies that the technologist should have on hand is presented in Table 1.1. This is not meant to be an exhaustive list but should be a good start for the beginning technologist.

Setup The type and site of surgery being performed will determine the type of monitoring that is needed. Many centers will have protocols that specify which modalities are to be monitored for each type of case. If a new type of procedure is being performed, the surgeon, anesthesiologist, and neurophysiologist should jointly decide which monitoring modalities are needed. Ideally, this should happen well before the morning of the case.

TABLE 1.1 A List of Supplies that Must be Present in the “Tool Box” of a Monitoring Technologist Supplies

Applications

A variety of tapes

To attach surface and needle electrodes

Conductive gel

To use for disc electrode conduction

Exfoliant

To prepare the skin prior to application of surface electrodes

Alcohol preps

To disinfect skin before placing needles and also for cleanup afterwards

Skin-marking pen

Head and muscle marking

Collodion

For gluing surface electrodes to head

Acetone

To remove collodion

1-inch gauze

Used with collodion to reinforce the attachment of surface electrodes

Conductive paste

To use for disc electrode conduction

Cotton applicator tips

To apply exfoliant

Syringes

To hold and apply collodion and conductive gel

Tape measure

To measure head and body markings

Air dryer

To dry collodion

Cotton balls

Used with acetone to remove collodion

Blunt-tip applicator

Used with syringe to apply conductive gel into disc electrode hole

Husain 01

1/17/08

11:51 AM

Page 15

CH AP T ER 1: Introd uction to the O pe rating R oom

Access to the patient should be coordinated with the surgical team, anesthesia team, and the circulating nurse. If there are surface electrodes that need to be applied, it is often easier to apply these in the preoperative holding area or induction room rather than in the operating room. For the patient’s comfort, needle electrodes are often placed after the patient has been anesthetized. It is important to place the electrodes securely, since it will be difficult to access them once the patient is prepped and draped. The wires should be guided along the patient’s body, toward the NIOM machine. They should be secured along the operating table and not left to dangle, so that they do not trip the surgeons or get caught in a C-arm or other piece of equipment. The setup is extremely important. It will determine the integrity of the waveforms and the ease with which the technologist can monitor the case.

Anesthetic Considerations Detailed descriptions of anesthetic applications are given elsewhere in this text. However, it should be mentioned that the monitoring technologist should have frequent, open, and clear communications with the anesthesia team prior to and during the case. A case requiring electromyographic (EMG) monitoring calls for few changes on the part of the anesthesiologist other than restricting the use of neuromuscular blocking agents. Cortical potentials of the SEPs are sensitive to inhalational agents and should be limited to a low and steady concentration if used; subcortical and peripheral SEP responses are less sensitive to these agents. Brainstem auditory evoked potentials (BAEPs) are resistant to most anesthetic agents, so there are few restrictions when monitoring only BAEPs. However, BAEPs are often monitored concurrently with cranial nerve VII (CN VII) EMG during posterior fossa tumor resections. In these cases, neuromuscular blocking agents should be limited. Indeed, many cases involve multiple modalities monitored together.

•

15

MEP monitoring is sensitive to inhalational agents. Additionally, neuromuscular blocking agents can obliterate MEP responses. When MEPs are performed, consideration should be given to a TIVA technique, as with propofol and opioids. EEG monitoring is sensitive to sedativehypnotic agents such as barbiturates, propofol, etc. These agents can induce a burst-suppression pattern. Low dose of inhalational agent may be preferred when EEG monitoring is being performed. A more complete discussion of anesthetics and their affect on NIOM is presented in a later chapter.

Remote Monitoring Many times the neurophysiologist who is responsible for interpreting the NIOM is not present in the operating room. In these situations, the neurophysiologist must review the data remotely. It is the monitoring technologist’s responsibility to connect the NIOM equipment to a remote monitoring system and ensure its working order. The neurophysiologist must do the same with his or her equipment. It is a good idea for the technologist to have a dedicated telephone line rather than using the operating room phone. The various systems available for remote monitoring are discussed in a later chapter.

SAFETY Safety is of utmost importance in the operating room. The monitoring technologists must be not only concerned with the safety of the patient but also with his or her own safety and the safety of the other personnel in the operating room. Attention to the issues listed below can help ensure safety for all.

Cords and Cables It has already been mentioned that electrode wires need to be neatly strung along the

Husain 01

1/17/08

16

•

11:51 AM

Page 16

SE C T I ON I : B a s i c P r i n c iple s

side of the bed so that the staff will not trip over them. There are many other cords and cables that the monitoring technologist must deal with. All of these should be run from the bed to the monitoring computer flat along the floor, neatly tied together. Some “no trip” operating room supplies are available to cover these cords for everyone’s safety.

Grounding The patient will have an electrical ground through the electrocautery system. An additional ground will be necessary for NIOM only if the monitoring equipment requires it as a system reference.

Personal Protective Equipment Personal protective equipment is worn by the monitoring technologist and other personnel in the operating room. If a CO2 laser is in use, everyone in the operating room should wear proper goggles to protect their eyes. When overhead ultraviolet (UV) lights are in use, a visor or plastic goggles should be worn to protect against UV light. During x-rays, including those taken with the C-arm (fluoroscopy), the technologist should either exit the room or wear lead protection covering at least the reproductive organs and thyroid. Gloves should always be worn when handling a patient, as this is fundamental to universal precautions.

Infection Control The NIOM team should always practice universal precautions for infection control. If needle electrodes are being used, the skin should be prepped with alcohol skin prep. Needle electrodes are for a single use only and should be disposed of in a sharps container. Surface electrodes may be reused and should be cleaned and disinfected using the individual laboratory’s protocol.

ETIQUETTE AND PROTOCOL As is true when one goes into any new environment, going into the operating room requires knowledge of etiquettes and protocols that are specific to the operating room. The monitoring technologist should familiarize himself or herself with these etiquettes and protocols so that the surgery can be helped to move along faster.

Burns Although burns are rare, they do occasionally occur and are most often associated with improper electrocautery grounding. When the electrocautery cannot find proper ground, it seeks another outlet for the current. The next available outlet is usually a monitoring electrode. Needle electrodes cause greater injury than surface electrodes because they are smaller and have a greater concentration of current in a smaller area. To prevent or reduce these types of injuries, the technologist should assist the circulating nurse in placing the electrocautery ground to ensure its proper placement. If a burn occurs, the surgeon and the neurophysiologist should be notified immediately.

Patient Access Access to the patient before a case begins can be somewhat of a competition. The preoperative nurse must check the patient in, get vital signs, place an intravenous line, and perform a variety of other tasks before the patient can go to the operating room. The circulating nurse must meet the patient, check allergies, etc. The anesthesia team must explain anesthetic protocol to the patient, check allergies, answer questions, and perform other tasks before anesthetic procedures are begun. The surgeons need patient access to mark the surgical site, answer last-minute questions, sign consents, etc. The monitoring technologist must meet the patient, explain his

Husain 01

1/17/08

11:51 AM

Page 17

CH AP T ER 1: Introd uction to the O pe rating R oom

or her part of the procedure and why it is important, take a history, and place electrodes. All this must happen in a coordinated manner so that an inordinate amount of time is not spent in this process. The monitoring technologist will likely find it easier to work with other members of the operating room team in a collaborative effort. In other words, it is most effective to introduce oneself to the other team members, explain what needs to be done, and ask when would be a good time to apply electrodes.

Attire Before entering the operating room restricted area, the NIOM staff must be dressed in the proper hospital-issued scrubs. They are usually supplied in the locker room or by vending machines. If scrubs are not available, disposable paper scrubs or “bunny suits” should be worn over street clothes prior to entering the restricted area. Head covering and shoe covers must be worn into the restricted area. These covers and disposable paper scrubs must be removed when leaving the restricted area. Face masks must be worn when entering the operating room.

Cleanup and Follow-up The work of the NIOM team does not end with the termination of the surgery. A substantial amount of time is required for disposing of supplies and cleaning others as well as performing follow-up of the service provided.

Disinfecting Equipment and Supplies Upon completion of the surgery, cables, head boxes, stimulating boxes, and amplifiers should be wiped down with a disinfecting solution and stored neatly on the NIOM machine. Any pencils, tape measures, or other tools that touched the patient or may have been cross-contaminated should be disinfec-

•

17

ted also. Any reusable electrodes should be scrubbed of any gross contamination and then soaked in a disinfectant. Disposable supplies should be disposed of properly.

Postsurgical Testing When the patient awakens from surgery, the monitoring technologist should be present for testing of the pathways that were monitored during the case. This includes moving toes, moving hands, feeling touch sensation on the feet and hands, smiling, hearing, or other testing as appropriate. This is especially important if changes were seen during NIOM, as they can be correlated to the neurologic examination. It is important for the technologist to be present for this testing so that, if the patient does not respond as expected, the surgeon can question the technologist (and neurophysiologist) about the NIOM. In addition, if the surgeon finds it necessary to reoperate emergently, the NIOM team may be asked to monitor for the emergent procedure.

DOCUMENTATION Thorough and complete documentation is of critical importance in NIOM. It serves not only to provide information about what happened during the casebut also to protect the NIOM team should postoperative complications occur (5).

Preoperative Documentation Prior to the date of surgery, if a preoperative EP or EEG study is performed, it must be interpreted and abnormal results communicated to the surgeon. On the day of surgery, the patient’s history must be recorded, with particular attention focusing on any contraindications for MEP monitoring. If such a contraindication exists, it must be discussed with the surgeon and noted in the patient’s medical record.

Husain 01

18

1/17/08

•

11:51 AM

Page 18

SE C T I ON I : B a s i c P r i n c iple s

Documentation During Surgery This is the most important documentation performed by the monitoring technologist. The exact modalities being tested and when each test is performed should be noted, along with the measured latencies of the peaks of interest or intensity of current needed to stimulate a particular structure. Many modern NIOM machines allow the technologist to enter notes directly onto the waveforms, making documentation easier. The technologists should also note the anesthetics used, their concentrations, and when they were changed. Physiologic parameters, such as blood pressure and temperature, and important surgical milestones should also be noted. When an alarm based on changes seen in NIOM is sounded, that must be recorded by the technologist. Preferably, the surgeon’s response to the alarm should also be noted. Similarly, discussions with the anesthesiologist should be noted as well. These help protect the NIOM team, as the documentation confirms that the NIOM team attempted to alert the responsible parties when problems occurred. As important as complete and thorough documentation is, noting irrelevant comments in the patient’s permanent medical record is not advisable. For example, a discussion between the attending surgeon and the surgical resident about how easy or difficult a particular case is does not need to be noted by the monitoring technologist. Arguments between operating room staff should also not be noted unless they are directly relevant to the NIOM.

Postoperative Documentation After the surgery is over, the monitoring technologist should make certain that documentation is complete. If certain elements of documentation could not be completed during the case because of concurrent responsibilities, they should be finished at this time. Significant findings noted on the postsurgery examination should also be noted. Per laboratory protocol,

data sheets (or waveform sheets/discs) are submitted to the neurophysiologist for report generation and storage. Unlike other types of reports for medical procedures, this NIOM report serves mostly a billing and medicolegal purpose rather than a clinical one. The clinical interpretation is provided in real time and not with the report. As such, the NIOM report should include the reason for NIOM, the modalities performed and how, whether any significant changes were noted and if so, what they were. Documentation that the surgeon was notified of these changes immediately is very important. Finally, the number of hours spent for interpretation should be included. The patient’s chart, along with technical (number of hours spent on the case by the monitoring technologist) and professional (number of hours spent interpreting the case by the neurophysiologist) charges are sent to the proper personnel.

CONCLUSIONS Performing NIOM is not simply performing clinical EP or EEG studies in the operating room. It requires a different knowledge and mind set. The operating room can be an intimidating environment for new NIOM staff. Beginners should learn the etiquettes and policies of their operating rooms. They must know who the other members of the operating team are and their roles. There are many pieces of equipment that may be present in the room, and the technologist must know which one will cause problems with NIOM. Attention to detail in setup, performing the NIOM, and afterwards will help the NIOM team provide a safe and effective service that will help reduce morbidity.

REFERENCES 1. Fortunato NH. Berry & Kohn’s Operating Room Technique, 9th ed. St. Louis: Mosby, 2000.

Husain 01

1/17/08

11:51 AM

Page 19

CH A P T ER 1: Introd uction to the O pe rating R oom 2. Turner S, Wicker P, Hind M. Principles of safe practice in the perioperative environment. In: Hind M, Wicker P, eds. Principles of Perioperative Practice. Edinburgh: Churchill Livingstone, 2000:17–50. 3. AORN. Recommended practices for perioperative nursing: section 3. In: Fogg D, ed. Standards, Recommended Practices, and Guidelines. Denver: AORN, 2004:209–399.

•

19

4. MacDonald DB. Safety of intraoperative transcranial electrical stimulation motor evoked potential monitoring. J Clin Neurophysiol 2002;19:416–429. 5. Nuwer MR. Regulatory and medical-legal aspects of intraoperative monitoring. J Clin Neurophysiol 2002;19:387–395.

Husain 01

1/17/08

11:51 AM

Page 20

Husain 02

1/17/08

11:52 AM

2

Page 21

Basic Neurophysiologic Intraoperative Monitoring Techniques

Robert E. Minahan Allen S. Mandir

monitoring is most effective in cases where nerve injury results from repetitive mechanical irritation of a nerve.

T

his chapter discusses and introduces the basic techniques of neurophysiologic intraoperative monitoring (NIOM). Techniques for electromyography (EMG), motor nerve conduction studies, somatosensory evoked potentials (SEPs), motor evoked potentials (MEPs), brainstem auditory evoked potentials (BAEPs), and electroencephalography (EEG) are all discussed. A full discussion of all signal acquisition parameters is beyond the scope of this chapter, but Table 2.1 summarizes the data necessary to allow proper configuration of these tests. All recording parameters are selected based on the expected frequency range, size, and latency of target signals while stimulus parameters are provided.

Anatomy and Physiology The most common surgeries in which EMG is monitored are those that place cranial nerves or spinal roots at risk. Therefore it is important that muscles innervated by the elements at risk be assessed. Commonly used spinal root or cranial nerve innervated muscles are listed in Table 2.2. Note that each spinal root innervates many muscles (and this group of muscles is termed the myotome for that root); conversely, most muscles are innervated by multiple spinal roots. In order to understand EMG potentials due to surgical irritation, we need to first consider the motor unit potential (MUP). The MUP is a group of muscle fibers innervated by a single axon. A single axon may innervate as few as three muscle fibers (as in eye muscles) or more than 500 (as in the gastrocnemius). In the typical scenario when surgical irritation of axons is sufficient, axonal depolarization results in the activation of the muscle fibers innervated by those axons. Depolarization of a single axon leads to single MUP, which is recorded as a “spike” on EMG. For NIOM, EMG is typically presented both visually on a monitor and aurally over a

FREE-RUNNING EMG AND MOTOR NERVE CONDUCTION STUDIES Background EMG can be monitored in any muscle accessible to a needle, wire, or surface electrode. Mechanical irritation of peripheral nerves or nerve roots results in muscle activity of the corresponding musculature. The consequent EMG recordings provide essentially instantaneous feedback to the surgeon regarding the effects of his or her actions. EMG 21

EEG

BAER 5000

1500–3000

50–70

500–1500

250–1000

500–1500

250–1000

1500

2000–5000

0–10,000

0.3–3

0–100

0–3

0.5–5

0.5–3

0.5–5

1–30

25–2000

10–150

20–1000

n/a

High

3–7

1.5–10

n/a

27–35

35–45

11–16

17–23

5–20

15–75

10–25

Low

*Values may fall outside these ranges in atypical, pathologic, or pediatric cases † Pedicle screw testing at durations other than 0.2 ms change the threshold ranges noted in the text

20

1 30–100

SEP–tibial nerve subcortical

Neuromuscular junction testing

30 30

SEP–tibial nerve cortical

30 30

SEP–median nerve subcortical

30

Motor evoked potentials (D-waves)

2000–5000 2000–5000

High n/a

High

20 mA–75 mA

75 dB–110 dB

n/a

25 mA–50 mA

25 mA–50 mA

20 mA–35mA

20 mA–35mA

100 V–500 V

200 V–1000 V

0–50 mA

Low

0.2–1

0.1–0.1

n/a

0.2–1

0.2–1

0.2–0.5

0.2–0.5

0.05–1

0.05–1

High

2–3

5–15

n/a

1.3–4.7

1.3–4.7

1.3–4.7

1.3–4.7

Manual–1

n/a

n/a 1–4

Low

n/a

High

Stimulation rate (Hz)

0.2–0.2†

Low

Stimulation duration (ms)

11:52 AM

SEP–median nerve cortical

10–40

Motor evoked potentials (myogenic)

10 10–20

Lumbar pedicle screw testing

Low

Stimulation intensity*

1/17/08

Free-running EMG

High Freq.

Typical latency (ms)*

•

Low Freq.

Typical amplitude (µV)*

22

Filters (Hz)

TABLE 2.1 Recording and Stimulating Parameters

Husain 02 Page 22

SE C T I ON I : B a s i c P r i n c iple s

Husain 02

1/17/08

11:52 AM

Page 23

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

•

23

TABLE 2.2 Muscles Commonly Used in EMG Monitoring* Cranial nerve–innervated muscles III, IV, VI – Extraocular muscles

T1 –Hand intrinsic muscles, flexor carpi ulnaris

V – Masseter, temporalis

T2,T3, T4, T5, T6 – Intercostal muscles, paraspinal muscles

VII –Frontalis, orbicularis oculus, orbicularis oris, mentalis, others

T6,T7, T8 – Upper rectus abdominis, paraspinal muscles, intercostal muscles

IX – Stylopharyngeus

T8,T9, T10 – Middle rectus abdominis, paraspinal muscles, intercostal muscles

X – Pharyngeal and laryngeal muscles XI – Sternocleidomastoid, trapezius XII – Tongue Spinal root myotomes C1 – None C2 – Sternocleidomastoid C3 – Trapezius, sternocleidomastoid C4 – Trapezius, levator scapulae C5 – Deltoid, biceps C6 –Biceps, triceps, brachioradialis, pronator teres, flexor carpi radialis (FCR) C7 –Triceps, pronator teres, FCR, forearm extensors C8 –Triceps, ulnar forearm muscles, all hand intrinsic muscles (incl. abductor pollicis brevis, first dorsal interosseous, adductor digiti minimi)

T10, T11, T12 – Lower rectus abdominis, paraspinal muscles, intercostal muscles L1 –Quadratus lumborum, paraspinals, cremaster ± iliopsoas ± internal oblique L2 –Iliopsoas, adductor longus, quadriceps, adductor magnus L3 –Quadriceps, adductor longus, adductor magnus, iliopsoas L4 –Quadriceps, tibialis anterior, adductor longus, adductor magnus, iliopsoas L5 –Tibialis anterior, peroneus longus, adductor magnus S1 – Gastrocnemius, abductor hallucis S2 – Gastrocnemius, abductor hallucis S2–S5 – Anal sphincter, urethral sphincter

*This list should not be considered inclusive and additional appropriate muscles for monitoring can be found in any diagnostic EMG text.

speaker. Irritation triggers motor units in a variety of patterns that are influenced by the preexisting condition of the nerve, the degree and mechanism of neural irritation, and the integrity of distal neuromuscular function. An EMG “burst” describes a brief period of polyphasic EMG activity representing the near simultaneous activation of multiple axons (motor units). An EMG “train” describes repetitive firing of one or more motor units lasting from a second to minutes. With intense multiaxonal irritation or voluntary muscle activation, MUPs may fill the recording channel. In this case, MUP overlap and no individual MUPs are distinguishable. When this is seen during diagnostic EMG, it is termed an interference pattern, and in relation to a surgical irritation can be thought of as an

extended burst. Typically this activity will cease with removal of the irritating factor, though a train of one or more MUPs may persist, as described in more detail below (Figure 1.1A to D).

Characterization and Interpretation In order to understand what level of clinical significance a pattern of EMG activity represents, the activity must be characterized beyond a simple burst or train description. The most important feature suggesting significance is a relation to surgical events at the time. In addition, a number of electrical features of EMG activity can suggest greater or lesser degrees of irritation and therefore greater or lesser clinical significance.

Husain 02

24

1/17/08

•

11:52 AM

Page 24

SE C T I ON I : B a s i c P r i n c iple s

A

B

C

D

E

FIGURE 2.1 EMG may be seen in a variety of patterns. A. A minor burst of activity occurring as a lumbar root is manipulated. B. A more intense burst occurring on the background of an ongoing train of activity. C. Intense ongoing trains of activity from multiple motor units (asynchronous activity). D. A residual train of activity as the effect of nerve root irritation wanes. E. An interference pattern in the left gastrocnemius muscle after inadvertent trauma to the corresponding nerve root.

Relation to Surgical Events The onset of EMG activity with a surgical action suggests a causative role. In addition to mechanical irritation, temperature (cold saline, heat from electrocautery) and osmotic irritation may induce intense EMG activity.

Recognition of nonmechanical causes is important, because under usual circumstances these have little or no clinical implication. Mechanical irritation, on the other hand, is associated with a risk of injury to the corresponding nerve either immediately or with

Husain 02

1/17/08

11:52 AM

Page 25

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

repetitive trauma. Mechanical irritation may induce a wide range of EMG patterns, as described in greater detail below; as a rule, however, the degree of irritation will correlate roughly with the intensity of the EMG activity. In addition, EMG activity that persists after the cessation of the irritative maneuver also suggests a relatively intense initial irritation. EMG activity may occur without apparent modulation related to ongoing surgical activity. The most clinically important but fortunately least common situation in which this might occur is when irritation is due to prior surgical actions that leave the nerve in an irritated state. Examples of such situations include a bone fragment that is in contact with a nerve or a nerve root compressed during tightening of instrumentation. If the causative role is not or cannot be identified immediately, ongoing EMG activity is likely to persist and the source of irritation may remain unidentified. More commonly, EMG activity that is not correlated to surgical events may be attributed to benign causes, such as return of muscle tone and voluntary muscle activity. Muscle tone may return with low levels of neuromuscular blockade and low anesthetic depth. Some muscles are more prone to demonstrate return of tone (e.g., frontalis, anal sphincter), but depending on the patient, this may include other muscles. In the extreme case of low anesthetic depth, voluntary muscle contraction may occur and electrical activity typically precedes gross clinical movement. Finally, EMG activity may be present prior to incision in some patients. Although a number of clinical conditions may lead to continuous muscle activity, in the setting of EMG NIOM, the cause is likely to relate to the reason for surgery (e.g., radiculopathy) and in these patients only increased EMG activity over this baseline is likely to reflect further nerve irritation. EMG activity may also be correlated to surgical activity when the surgical activity is trivial or remote from the neural elements

•

25

activated. Once some degree of nerve damage is established (either intraoperatively or preoperatively), minor manipulation of the corresponding nerve or even patient movement may elicit strong EMG activity (see below). For example, in many patients with severe preoperative spinal stenosis, EMG activity may be elicited by trivial surgical actions due at least in part to an increased proclivity of axons to fire. EMG activity in these cases is common and, when associated with surgical actions remote from neural elements, poses a very low risk of further damage to nerves. Finally, it is possible that reflexive EMG activity may be triggered in myotomes remote from surgical activity through nociceptive input or other mechanisms. Such reflexive mechanisms are not well described in the EMG monitoring literature; however, a clear correlation of EMG activity to surgical activity, even when not in the expected myotomes, should trigger suspicion of a causative role.

Number of MUPs Each distinct MUP represents a separate depolarizing axon. Thus, greater numbers of axons affected by any irritative source will result in larger numbers of distinct MUPs recorded and a greater intensity of recorded EMG activity.

Firing Rate The rate at which MUPs fire correlates to the degree of irritation, with higher rates indicating a greater degrees of irritation. Rates of firing for individual units tend to be constant over short periods but may gradually wax (persistent irritation) or wane (irritative source typically removed) over time. Activity may abruptly cease at any time. Most firing patterns are regular for individual MUPs; but when many MUPs are activated, each will have its own onset, offset, and rate. This combination of different rates, initiation, and cessation of MUPs results in what has been described as an “asynchronous” pattern of EMG firing. The extreme case of asynchro-

Husain 02

26

1/17/08

•

11:52 AM

Page 26

SE C T I ON I : B a s i c P r i n c iple s

FIGURE 2.2 Repetitive microtrauma to nerves may produce intense EMG activity, as seen in this vestibular neuroma resection. The orbicularis oculi shows myokymia and the orbicularis oris shows asynchronous high-rate trains of activity reflecting significant irritation and likely clinically apparent postoperative neuropathy.

nous firing is the interference pattern, as described above (Figure 2.2).

Shape The shape of MUPs is largely determined by the distance and orientation of the MUP with respect to the recording electrodes coupled with the number of muscle fibers in the MUP. Some authors have tried to correlate the shape of the MUP with significance (multiphasic being more significant than monophasic), but this distinction is likely to be most useful in distinguishing artifact from MUP. When correctly identified, a monophasic MUP is likely to reflect recording from a distant muscle. The monophasic MUP therefore suggests ongoing axonal depolarization, although the absence of closer units may suggest relatively sparse activity. In addition, the shape and size of any single MUP is usually consistent over time. However, with partial neuromuscular blockade, some variation may be present due to inconsistent activation of component muscle fibers.

Amplitude High-amplitude EMG activity resulting from the superimposition of multiple MUPs

suggests a higher degree of irritation, as discussed above under “Number of MUPs.” On the other hand, the amplitude of any single MUP is likely unimportant, as a single MUP represents activation of a single axon. No studies have determined if the axons associated with either large or small MUP are preferentially activated due to mechanical irritation; therefore it may be that small or large units are essentially recruited by chance. In addition, chronic nerve or root denervation may lead to remodeling of motor units, with more muscle fibers innervated by single axons. If this is the case, patients with preexisting chronic radiculopathy are likely to have large motor units. Finally, the recorded amplitude is also a function of how close the motor unit’s muscle fibers are to the recording electrode.

Relation of EMG Activity to Degree of Nerve Dysfunction EMG is the only signal in NIOM in which an absence of activity is the expected state in both normal nerves and with complete loss of function of the monitored neural target. Given this convergence of extremes, one must examine the effect of a given irritative maneuver through the entire spectrum of nerve dysfunction. Stripping of epineurium, perineurium, and possibly myelin appears to increase the propensity of axons to depolarize in response to irritants. As a result, the undisturbed nerve is less likely to be irritated with minor manipulation than a nerve that has undergone nonaxonal damage. Once the axon itself is disrupted, that axon will not produce MUPs with proximal irritation (though the cut end of the distal portion of the axon may retain an ability to depolarize). Thus, as progressive axonal injury occurs, the nerve may once again show lesser degrees of EMG activity with manipulation proximal to the injury. This phenomenon is best illustrated when a nerve is subject to repetitive microtrauma, as with resection of neural tumors. In these cases, first, progressively increasing intensity

Husain 02

1/17/08

11:52 AM

Page 27

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

•

27

of EMG activity is often seen. At its maximum, myokymic potentials and very high rates of neurotonic trains may be seen (50 to 100 Hz) (1). When myokymic or high-rate neurotonic patterns are encountered, the gradual subsequent reduction in EMG activity despite continued surgical manipulation may reflect either less irritation or suggest progressive dysfunction of the nerve (Figure 2.2). Repetitive microtrauma to the nerve nicely demonstrates the spectrum of EMG responses that might be observed with gradual and progressive dysfunction. However, the mechanism of neural injury will determine which if any of the intermediate stages of dysfunction are reached. For example, abrupt transection of a nerve may be associated with a single burst of EMG activity followed by electrical silence thereafter. In fact, sharp transection may not have any EMG correlate. As a result, one must consider the mechanisms of potential injury and the likely timing of associated progression through the range of dysfunction in considering the significance of EMG in relation to the surgical activity at the time (Figure 2.2). The onset of nerve ischemia may be irritative under appropriate conditions, but under most conditions we consider ischemia to be nonirritative and therefore no EMG may be present. As a result, compressive nerve insults may yield variable EMG correlates if coincident ischemia blocks axonal depolarization. In addition, any source of neuromuscular dysfunction distal to the site of surgical irritation will reduce sensitivity from EMG monitoring and allow false-negative results. These include distal nerve compression or ischemia, distal neuropathy, limb ischemia, excessive neuromuscular junction blockade, or inexcitable myopathy.

plexities involved with EMG monitoring. It is the monitoring team’s duty to identify the location, pattern, intensity, and duration of this activity and then communicate it to the surgeon when appropriate. At least as important as the type of EMG activity is the relation to surgical activity at the time of occurrence. Thus, the neurophysiologist must maintain awareness of surgical events and give feedback immediately upon occurrence, so that the surgeon understands the relationship of EMG to his or her activities. Short bursts of activity and low-frequency trains of activity are usually, at worst, lowlevel irritation with a low risk for persisting injury. The burst is typically a physiologic phenomenon due to the excitation of mechanoreceptors on the axon. The report of brief bursts to the surgeon should suggest manipulation of a nerve but little else. The surgeon may find reports of bursts helpful if he or she is not aware of being in contact with neural structures or if that activity is restricted to simple bursts without more ominous patterns present. Such activity is often described as “minor,” so that its typically benign nature is clear. Activity that persists after the cessation of the offending surgical action suggests some level of ongoing insult or injury (often subclinical) and is more often described as “significant.” Persistent trains of one or a few MUP at lower rates are described as moderate activity, and activity with many MUP or particularly high train rates described as intense.

Communication with the Surgeon

Spinal instrumentation must be anchored to the vertebral column in order to provide support and allow bony fusion. Screws are often placed in the pedicle to provide this anchor; if they are malpositioned, however,

EMG is considered by many to be the simplest form of NIOM. While this may be the case, the above discussion shows that one must still have a healthy respect for the com-

Motor Nerve Conduction Studies (“Triggered EMG”) Electrical Evaluation of Pedicle Holes and Screws

Husain 02

1/17/08

28

•

11:52 AM

Page 28

SE C T I ON I : B a s i c P r i n c iple s

they may impinge on exiting nerve roots, causing radiculopathy. Pedicle holes and screws should be electrically tested to assess for perforation of the pedicle wall. If the hole drilled in the pedicle perforates the wall, a low-impedance pathway will be created between stimulation within the hole and nearby exiting nerve roots. When a perforation exists, a relatively low level of electrical stimulation will activate these nerve roots, with resultant activity recorded from their corresponding muscles. Thus the integrity of the pedicle can be assessed based on the minimum level of electrical current needed to activate nearby nerve roots. At lumbosacral root levels using monopolar cathodal stimulation at 0.2-ms duration, thresholds can be interpreted as shown in Table 2.3 (2). Typically, if pedicle screws stimulate (activate nearby roots) at less than 7 mA, a perforation has occurred, whereas if they stimulate at greater than 10 mA, a perforation is unlikely. For holes, however, if activation is noted at less than 5 mA, a perforation is likely; if they stimulate at greater than 7 mA, perforation is unlikely. In this context, “hole” refers to stimulation of the pedicle hole using a ball-tipped electrode as the cathode. If, instead, the hole is indirectly stimulated via an instrument (e.g., tap), then current shunting may occur if that instrument is in contact with any tissue or fluid and results will be skewed. “Screw” refers to stimulation of the head of the screw shaft after placement. If, instead, the mobile top of a polyaxial screw is stimulated, there may be an inconsistent electrical connection to the shaft and again results may be skewed. In cases where low thresholds are TABLE 2.3 Threshold Values Indicating the Likelihood of Pedicle Screw Malpositioning Perforation probable

Perforation possible

Perforation unlikely

Hole

7 mA

Screw

10 mA

found, the surgeon may choose to remove or redirect the screw at that site. Alternatively, in a situation where redirection is not practical or a screw is particularly important to the success of fusion, the surgeon may leave the screw in place after probing the hole and/or performing fluoroscopy or other radiographic imaging in an attempt to determine if the pedicle wall is likely to be perforated in a clinically significant manner—i.e., there is potential impingement on a nerve root. The presence of free-running EMG activity corresponding to screw placement or probing suggests that a perforation with nerve impingement is present. Pedicle screws or lateral mass screws at thoracic and cervical levels may also be electrically evaluated, and cutoff threshold values are likely to be similar to those for lumbosacral levels. However, evidence for specific values remains to be fully determined (Figure 2.3).

Motor Nerve Conduction Studies for Purposes Other than Pedicle Screw Evaluation Motor nerve conduction studies may be used in a variety of settings where motor nerves are accessible to the surgeon and potentially at risk of injury. In these cases, the surgeon will directly stimulate nerves of interest and motor responses recorded in the corresponding muscle. This technique may assess the continuity and function of nerves from the point of stimulation to the muscle and can explore whether neural structures are in or near an area of proposed resection or electrocautery. A common use of these techniques is for vestibular schwannoma resection, where the facial nerve must first be located and then assessed.

Anesthesia Anesthetic agents have little effect on the recorded muscle responses of EMG or motor nerve conduction studies. Neuromuscular

Husain 02

1/17/08

11:52 AM

Page 29

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

•

29

FIGURE 2.3 A CMAP is seen in the right quadriceps muscle after stimulation of the right L3 pedicle screw at 3 mA. This is a low threshold, indicating a low-impedance pathway (pedicle wall perforation) between the screw and exiting nerve roots.

blocking agents, on the other hand, may have a profound effect if sufficient residual neuromuscular function is not present. When partial neuromuscular blockade is used during EMG monitoring or motor nerve conduction studies, we target a level that does not depresses the abductor pollicis brevis compound motor action potential amplitude more than 80% relative to an awake value. This level is reasonably approximated by a train-of-four ratio (ratio of the fourth to first compound motor action potential using repetitive nerve stimulation at 2 Hz) in the gastrocnemius muscle of > 0.3 or the presence of at least two responses in the abductor pollicis brevis exceeding 600 µV in amplitude (unpublished data). If the anesthesiologist is assessing visual responses to train-of-four testing, then four out of four twitches is the only reliable criterion due to the imprecision of that test. However, in most cases, two or three visible twitches will also be adequate, but the confidence in this finding is reduced.

SOMATOSENSORY EVOKED POTENTIALS Background SEPs have been the primary spinal cord monitoring modality for decades and have more recently been augmented but not supplanted by motor evoked potentials (MEPs). SEPs remain a ubiquitous form of neuromonitoring because they assess all levels of the neuraxis from the peripheral nerve to the cerebral hemispheres. SEPs are well suited for NIOM for a number of reasons. First, with modern techniques and equipment, responses have a definable amplitude and latency that can be quantified for comparison throughout a procedure. Second, the signals have reasonable stability, so that injury can be identified with confidence. Third, multiple recording sites can be employed along the course of the assessed somatosensory pathways and the neural generators for each of these sites are known within practical precision. This latter aspect

Husain 02

1/17/08

30

•

11:52 AM

Page 30

SE C T I ON I : B a s i c P r i n c iple s

allows localization to peripheral nerve, spinal cord, or brain when a neural insult occurs and thus allows the most appropriate corrective actions to be taken. Finally, an SEP can be elicited from almost any nerve containing sensory fibers, with median, ulnar, and tibial nerves used most frequently.

SEP Generators and Localization Localization of neural deficits based on patterns of SEP changes relies on an understanding of the neural generators producing the observed waveforms. If the generators are known, it is a simple matter to understand that

after stimulation of a single nerve, loss of signals at a point along the course of its propagation to the brain suggests an insult between the most proximal generator with a retained signal and the next most proximal generator with a degraded signal. Table 2.4 shows generators for median and tibial nerve SEPs. The cervicomedullary signals generated by upper extremity SEPs are among the most complex, because multiple signal components contribute to the response recorded at 13 to 14 ms when using standard recording techniques. In recording with an electrode over the posterior neck, there are negative components contributed by the cervical spinal cord.

TABLE 2.4 Neural Generators for Median and Tibial Nerve SEP Generators Median Nerve SEP Generators

Tibial Nerve SEP Generators

Generator

Common channels used

Alternate labels

N9

Brachial plexus

EPi-EPc

Erb’s

N11

Spinal nerve root

Crv-Fpz

N13a

Dorsal horn interneurons

Crv6-Fpz

N13b

Dorsal column

Label

Label Popliteal

Generator

Common channels used

Alternate labels

Tibial nerve Popliteal action potential

N23

Dorsal horn interneurons

T12-iliac cr. Lumbar point

Cervical, subcortical

P31

Medulla

Crv-Fpz, Mast-Fpz

Cervical, subcortical

Crv2-Fpz

Cervical, subcortical

N34

Primary sensory cortex

Cc-Fpz

N37

P13

Spinomedullary Crv-Fpz, Mast-Fpz junction

Cervical, subcortical

P38

Primary sensory cortex

Ci-Fpz, Cz'-Fpz, Ci-Cc, Cz'-Cc

P39, P40, cortical

P14

Lemniscal paths, Crv-Fpz, cuneate nucleus Mast-Fpz

Cervical, subcortical

N38

Primary sensory cortex

Cc-Fpz

N18

Brainstem/ thalamic

Cinoncephalic

N19

Primary sensory cortex

Cc-Fz, Cc-Ci

P22

Primary motor cortex

Cc-Fz, Cc-Ci

N20, cortical

EP = Erb’s point; Crv = electrode over spinous process, Crv2 and Crv6 refer to the vertebral level; Cc = C3' or C4', whichever is contralateral; Ci = C3' or C4', whichever is ipsilateral; Bold = more important and more consistently recorded signals.

Husain 02

1/17/08

11:52 AM

Page 31

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

At roughly the same latency there are positive components recorded from Fpz that are contributed by the medulla. Referencing these electrodes together summates the component signals and produces the observed “subcortical” response. When a surgery is performed at the cervical level, it is best that the entire subcortical signal be generated above the level of surgery so that it will be reliably affected should a cervical-level cord injury occur. For this purpose a mastoid or ear electrode is referenced to FPz to record a signal that is generated at or above the foramen magnum. In this case a cervical-level electrode is omitted so that the associated spinal cord generators do not contribute to the recorded subcortical signal. This prevents a portion of the subcortical signal to persist in the presence of a cervical cord injury and avoids recording novel generators that may arise with an injury (far-field signals due to abrupt termination of axon conduction). As an example, an anterior cervical spine procedure may be associated with a number of different SEP changes reflecting different events (Table 2.5). Localization must also include information from other monitoring modalities. Figure 2.4 is a demonstration of bilateral loss of tibial nerve SEP signals during lower thoracic decompression and fusion in a patient with spinal stenosis due to Pott’s disease. The loss of cortical SEP signals following tibial nerve stimulation occurred in the setting of intact signals at the popliteal fossa, intact SEPs following median nerve stimulation, and intact MEP signals. These findings suggest dysfunction of large fiber somatosensory pathways

•

31

above the level of the knee and below the cervical cord. This was presumed to result from thoracic posterior column dysfunction at the level of surgery, and the surgeon elected to expeditiously decompress the spine at that level. Signals returned as decompression was achieved.

Interpretation The most commonly used criteria for identifying a significant degradation of the SEPs are drop in amplitude below 50% of the baseline and/or a latency prolongation of 10% over the baseline value. These criteria are based on clinical experience and standards may vary among NIOM groups. Application of these “standard” criteria may need to be modified in cases where signals are of poor initial quality or where confounding systemic factors exist, such as alteration of anesthetic gases, low level of neuromuscular blockade, dramatic blood loss, or other metabolic derangements. When a “significant” SEP degradation meets the above criteria, it must, of course, be interpreted within its clinical context. Once again, localization of the dysfunction allows appropriate action and may prevent warnings to the surgeon in cases where anesthetic, positioning, or technical issues are present. In addition, correct localization can help direct the surgeon to the most appropriate course to correct dysfunction due to surgical events. Finally, timing of the SEP changes in respect to surgical events is helpful in identifying a culprit surgical maneuver. This latter statement

TABLE 2.5 Localization of Neural Dysfunction Based on Pattern of SEP Changes* Locus of neural insult

Associated pattern of SEP degradation

Spinal cord dysfunction

Loss of subcorticaland cortical signals, Erb’s point intact

Limb malpositioning

Unilateral loss of Erb’s point, subcortical and cortical signals

Cerebral ischemia (carotid retraction)

Unilateral cortical loss, intact subcortical signal

Anesthetic effect

Global cortical loss, intact subcortical signals

*This allows identification of a number of different sources of insult during an anterior cervical procedure.

Husain 02

32

1/17/08

•

11:52 AM

Page 32

SE C T I ON I : B a s i c P r i n c iple s

A

B

FIGURE 2.4 Tibial SEP signals were lost during a posterior thoracic decompression and fusion in a patient with spinal stenosis due to Pott’s disease. The signals were lost early in the procedure (A) and recovered with decompression. The MEP signals (B) were stable at the time, suggesting an isolated posterior column effect.

Husain 02

1/17/08

11:52 AM

Page 33

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

comes with the caveat that the advent of MEP testing has made clear that SEP changes may be significantly delayed when dysfunction primarily affects the anterior spinal cord.

Indications Given that SEP assess all levels of the neuraxis, it is not surprising that SEPs are a staple part of most neuromonitoring configurations. Any surgery that places somatosensory pathways at significant risk is likely to benefit from monitoring with SEPs if a corrective action is possible when dysfunction is identified. A review of SEP monitoring at all levels of the neuraxis is beyond the scope of this chapter, but appropriate cases include cervical or thoracic-level spinal surgeries, posterior fossa surgeries, hemispheric and deep brain surgeries, and surgeries that place the peripheral nerve at risk. The SEP has poor sensitivity at the level of the spinal nerve root. SEPs from peripheral nerves are typically mediated by multiple nerve roots and, as such, have poor sensitivity to monoradiculopathies. Despite this, surgeries that place multiple roots at risk simultaneously (e.g., the cauda equina) may still benefit from SEP monitoring. The most common use of SEPs is for spinal cord monitoring; thus this deserves special note. A 1995 Scoliosis Research Society poll found that there was a 50% decline in major neurologic deficits associated with scoliosis surgery after introduction of NIOM, that SEP identified insults with a sensitivity of 92% and a specificity of 98.9%, and that participation of experienced monitoring teams was associated with half the rate of neurologic deficits compared to results with inexperienced teams (3).

Anesthesia Anesthetic considerations for SEP monitoring are discussed in greater detail in another chapter. As a general principle, volatile anesthetics suppress cortical SEP sig-

•

33

nals in a dose-dependent manner. On a “per MAC” basis, this effect is similar for nitrous oxide and potent inhalation agents; however, the rapid onset and offset of nitrous oxide’s effects when its concentration is altered make its suppressive effect potentially more difficult to distinguish from neural insults. Propofol also may suppress SEP signals, but with less potency than the volatile agents. Rigid cutoff values for anesthetic levels are not necessary for the management of SEPs, although guideline targets may be helpful to the anesthesia team. Levels must be sufficiently low that robust and stable signals are obtained at a postequilibration baseline. From that point forward it is the stability of the anesthetic regimen that is most important.

TRANSCRANIAL ELECTRICAL MOTOR EVOKED POTENTIALS Background Traditional NIOM of the spinal cord has relied upon SEPs despite the fact that motor dysfunction is the most feared form of spinal cord injury. SEPs allow monitoring of signals that traverse the dorsal columns of the spinal cord, whereas motor function is only indirectly assessed, relying on coexisting dysfunction in sensory pathways to predict motor injury. Unfortunately, motor pathways may be injured while sparing sensory pathways for at least two reasons. First, there is anatomic separation between motor and dorsal column sensory pathways, with motor structures residing in the anterior cord. Second, the vascular supply to the anterior spinal cord is distinct from that to the posterior columns. The anterior spinal cord may be selectively vulnerable to ischemia due to a less robust vascular anastomotic network that does not compensate as well as the posterior vasculature in the face of hypoperfusion. In addition and possibly more importantly, motor gray matter (neurons) in the spinal cord is far more sensi-

Husain 02

34

1/17/08

•

11:52 AM

Page 34

SE C T I ON I : B a s i c P r i n c iple s

tive to ischemia than the dorsal column sensory white matter (axons) and thus is selectively vulnerable for this reason as well. The lack of direct anterior cord monitoring with SEP is problematic, and cases of isolated injury with preserved SEP spinal cord monitoring have occurred. In addition, it is now clear that MEP signal changes may precede SEP signal changes, allowing earlier identification of spinal cord dysfunction and thus better correlation to surgical events with more timely intervention. At the same time, SEPs remain useful in identifying sensory dysfunction and in providing a second, independent test of cord function. In addition, posterior columns may also be injured while anterior cord pathways are spared (see Figure 2.4). Thus, MEPs are almost always performed in conjunction with SEP monitoring, and these test modalities function as complementary tests.

MEP Technique Recording Transcranial stimulation nonspecifically activates motor fibers subserving many or all levels of the spinal cord, and choosing appropriate muscles for recording is important to localize motor tract deficits. Ideally, recording is performed from at least two muscles on each side below the surgical level and one muscle above it to serve as a control signal. Common muscles used for this purpose are abductor pollicis brevis, tibialis anterior, and abductor hallucis. Other muscles may be added as needed for redundancy and greater localization precision (see Table 2.2 for commonly used muscles). Alternatively, averaged “D-wave” responses may be obtained from the spinal cord white matter tracts using an electrode placed in the epidural or subdural space. This technique allows high-quality signals without significant degradation from anesthetic agents or neuromuscular blockade. D waves are not

discussed in detail in this chapter but excellent references are available (4,5).

Stimulation Transcranial activation of subcortical motor tracts is elicited most efficiently with anodal stimulation. A number of electrode configurations may be used for transcranial stimulation, but all utilize anodal stimulation with the anode placed over or adjacent to the targeted portion of the motor strip. Although an optimized configuration is yet to be agreed upon, usually one electrode is placed on C1 (expanded International 10–20 System) and the other on C2, with the anode being on the side of the targeted brain and the other electrode used as cathode. In any given patient the optimal location of these electrodes is variable and more lateral placement at C3 or C4 for one or both of the electrodes may be attempted. MEPs are best elicited using a train of stimuli, which serves to overcome anesthetic inhibition of the synapse at the anterior horn cell of the spinal cord. Trains of four to nine stimuli are delivered with an interstimulus interval of 2 to 4 ms.

Signal Optimization Myogenic MEP signals are frequently variable in their amplitude and morphology, making interpretation of changes challenging compared to other NIOM methods. If MEP signals are poor, there are three general types of causes to consider: signal acquisition methods, anesthetic/systemic variables, and patient pathophysiology. Signal acquisition methods relate to factors controlled by the monitoring team and the monitoring system must be confirmed as working correctly with electrode positions and stimulation parameters optimized. Anesthetic and systemic variables are those factors managed by the anesthesiology team, such as anesthetic agents, blood pressure, temperature, oxygenation, neuromuscular blockade, and myriad others. Finally, patient-related factors may relate to physiologic or pathophysiologic

Husain 02

1/17/08

11:52 AM

Page 35

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

sources of poor MEP signals. Preexisting dysfunction of motor pathways such as stroke, myelopathy, or neuropathy may limit the initial quality of signals, while deterioration of MEP signals may reflect surgical injury to motor pathways or benign causes such as alteration of stimulation efficiency resulting from scalp edema or cerebrospinal fluid (CSF) drainage (Figure 2.5). Degradation of signals is often multifactorial, and adjustments or allowances must be made for factors within any or all of these general categories. A full discussion of all possibilities within these categories is beyond the scope of this chapter, but at minimum, each category should be considered in detail when MEP signals are suboptimal or deteriorating.

MEP Interpretation There are no universally accepted criteria for identification of significant myogenic MEP signal changes. While this may be a source of frustration for those looking for easy “rules” by which to interpret MEPs, it should come as no surprise. Initially, obtained MEP signals in one patient may be robust and stable, may be widely variable in their amplitude and morphology in another, and may be absent in

•

35

another despite an optimized regimen. For a given patient, initial signal quality may fall anywhere along the continuum from robust to absent. If excellent monitoring techniques are assumed to be used, the variability among different patients stems from varying anatomy, degrees of preexisting motor pathway dysfunction, and sensitivity to anesthetic agents used. More importantly, one should not expect to apply the same interpretative criteria to signals that are robust initially as to those that are initially barely present. A number of criteria have been proposed, including a specific percentage of signal loss, presence vs. absence of signals, an increase in stimulation threshold necessary to obtain signals, and complexity/ polyphasia of the recorded response. Any criterion must be applied in a manner that makes sense based on initial signal quality and variability as well within the context of changes in systemic variables that may occur. For robust signals, the authors use complete loss of a signal or abrupt significant decrease in amplitude of 80% or more in the absence of an explanation other than surgical injury. Gradual changes in signals more commonly reflect systemic factors or an “anesthetic fade” phenomenon, so gradual changes might be given less credence unless the onset

FIGURE 2.5 Poor-quality MEP signals may be due to suboptimal signal acquisition methods, unfavorable anesthetic or systemic variables, and/or the patient’s physiology or pathophysiology.

Husain 02

1/17/08

11:52 AM

Page 36

A

B

C

FIGURE 2.6 Right-body MEP signals are lost during scoliosis correction. MEP signals are robust immediately preceding spinal correction (A). Lower extremity right-body MEP signals are lost soon after correction was initiated, while the left-hand and left-body signals are preserved (B). Signals immediately started to improve after corrective forces were relaxed and were comparable to baseline minutes later (C). APB = abductor pollicis brevis, IL = iliopsoas, QU = quadriceps, TA = tibialis anterior, GA = gastrocnemius, AH = abductor hallucis. 36

Husain 02

1/17/08

11:52 AM

Page 37

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

of the change can be related to a surgical event that may result in gradual dysfunction (6). For poor initial signals, the criterion used is usually limited to the presence or absence of the signal. This criterion risks false-negative findings but represents a balance with what would otherwise be an untenable false-positive rate (Figure 2.6A to C).

MEP Complications and Contraindications The most common complication of MEP testing is tongue laceration due to contraction of the masseter muscle triggered by transcranial stimulation. This can be minimized by routine use of a soft bite block. Another near universal aspect of MEP monitoring is patient movement due to muscle contraction after transcranial stimulation. This is an expected part of MEP, although it could lead to complications if movement occurs unexpectedly during a delicate portion of a surgery. As a result, the surgeon is typically warned prior to stimulation, and testing is delayed if there is an objection from the surgeon. These and the following less common complications have been reviewed by McDonald (7), who reported one case of mandibular fracture, presumably due to masseter contraction. Seizures have rarely been seen intraoperatively in the setting of MEP testing, with the review by McDonald finding 5 instances in 15,000 cases. Cardiac arrhythmia was reported in one case, but there was no clear relation to the MEP stimulation. Interference with cardiac pacemaker programming is a theoretical concern, but no reports of interference have been made and the magnetic fields produced by MEP stimulation are likely to be small relative to routinely used electrocautery. The authors’ practice in the setting of a pacemaker is to proceed with MEP testing but to discuss this low risk with the anesthesia team, so that they are prepared should any cardiac anomaly arise. Contraindications to MEP testing include the presence of indwelling deep brain stimula-

•

37

tion electrodes and cochlear implants. In addition, the presence of plegia in the target limbs obviates the need to attempt MEP testing and makes successful acquisition unlikely. Finally, we avoid MEP testing in patients with recent craniotomy or skull fracture for two reasons. First, we do not want to cause damage by shifting unstable skull elements due to scalp muscle contraction, and second, we do not want to place electrodes into skull defects. There are few data on patients under 2 years of age, and similar skull stability concerns should be considered.

MEP Indications MEP monitoring adds sensitivity and confidence to more traditional SEP monitoring protocols. MEPs are used in those cases with a reasonable probability of motor injury that spares dorsal columns sensory pathways and/or in high-risk procedures where a second independent test of spinal cord function is desired. For spinal cord monitoring, some will argue that MEPs should be used in every case where the spinal cord is at risk, while others are still more selective. Although this debate cannot be resolved here, there is good evidence for addition of MEP monitoring in intramedullary spinal cord tumor cases, most cervicothoracic spinal deformity cases, thoracic aortic aneurysm surgeries, and complicated cervical-level cases. In addition, MEP monitoring may also be useful at other levels of the nervous system, as for selected intracranial tumors or cerebral aneurysms. Finally, the use of MEP for lumbar-level procedures is usually not recommended on a routine basis due to a low likelihood of disassociated sensory and motor injury and imperfect sensitivity of MEPs to individual root injury.

Anesthesia The stimulus for transcranial electrical MEPs activates subcortical white matter, which effectively bypasses the cortical effects

Husain 02

38

1/17/08

•

11:52 AM

Page 38

SE C T I ON I : B a s i c P r i n c iple s

TABLE 2.6 Anesthetic Agents Inhibit MEPs with Varying Levels Of Potency Degree of MEP inhibition

Anesthetic agent

High

Potent volatile agents, N2O

Moderate

Propofol, benzodiazepines

Little or none

Etomidate and derivatives, ketamine, narcotics

of anesthetic agents. Unfortunately, anesthetics are also potent inhibitors of the synapses at the anterior horn cell in the spinal cord, and this is the primary locus of the undesirable inhibition of MEPs. In a simple sense, one can stratify these agents according to the extent of their effect on the anterior horn cell, as in Table 2.6. Ideal anesthetic regimens for MEP monitoring are a matter of significant debate. However, grouping into potent, moderate, and little inhibition is widely accepted. As one shifts from more a potent inhibition to less potent group (at similar depths of anesthesia), MEPs will tend to improve (if originally present) or be more likely to appear (if originally absent). At some centers, those anesthetic regimens with low levels of inhibition may be less familiar to the anesthesia team and therefore more difficult to manage compared to standard anesthetic regimens. In general, the acquisition of robust and therefore reliable MEPs will be dramatically aided when anesthesiologists experienced in MEP anesthesia are available. Finally, lower inhibition regimens may also be slightly more expensive, but cost differences are minor. Ideally, anesthetic regimens can be adjusted and tailored to the individual patient based on MEP quality during the early portions of the procedure. However, given limited available time, this may not be feasible in all cases; in most cases, a gradual, stepwise shift will not allow adequate acquisition and evaluation of signals before critical portions of the procedure begin.

Neuromuscular blockade will also impact MEP signals. In general, neuromuscular blockade has a more predictable and linear effect on MEPs, whereas volatile anesthetics have an unpredictable nonlinear effect. Typically, at reasonable levels of partial neuromuscular blockade, MEPs will be obtainable if the anesthetic agents are appropriate. However, if only low-amplitude MEP signals can be obtained, it may be necessary to forgo the use of partial neuromuscular blockade.

ELECTROENCEPHALOGRAPHY (EEG) Background EEG is a measure of cortical activity in real time. As it is applied in NIOM, EEG often is useful to detect widespread, gross changes in cortical function and does not require the high spatial resolution typically found in diagnostic EEG studies. Since it is not an evoked potential, EEG cannot make use of averaging techniques to increase its signal-to-noise ratio. However, the recorded EEG is typically much larger (tens of microvolts) compared to evoked responses like SEPs.

Generators EEG recordings comprise postsynaptic potentials from cerebrocortical neurons. Surprisingly, cortical action potentials, though larger in amplitude than postsynaptic potentials, do not contribute to the EEG waveform due in part to physiologic filtering through tissue and bone as well as the longer duration of postsynaptic potentials. Furthermore, not all cortical neurons are represented in recorded EEGs from the scalp; it is the large pyramidal cells in cortex layer V that are primarily represented in scalp EEGs. This is due to the relatively large postsynaptic potentials produced in layer V of the cortex and the vertical dipoles created, which are more likely to be recorded at the scalp. Finally, patterns and

Husain 02

1/17/08

11:52 AM

Page 39

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

distribution of synchronous firing also favors these large pyramidal cells to be represented in surface EEGs.

Classification of Frequency Frequencies of EEG waveforms are classically divided into ranges as follows: delta < 4 Hz; theta 4 to 7 Hz; alpha 8 to 12 Hz, and beta 13 to 30 Hz. In normal individuals without the presence of anesthesia, states of wakefulness correlate to frequency, with the higher ranges expressed in states of wakefulness. During NIOM, however, when anesthetics are present, the depth of anesthesia will influence the EEG frequency. As a general rule, as anesthesia is induced, the EEG goes through a stage of activation (beta activity) early on, which then gives way to slowing. As the depth of anesthesia increases, slowing potentiates, which leads to burst suppression and eventually to electrocerebral inactivity.

Recording Montages The choice of EEG montage may need to be tailored according to the surgery at hand. Reasons for modifying the montage may result from physical limitations in electrode placement or the need to focus on a region of interest. For intracranial surgeries, the location of the sterile field may necessitate adjustment of a standard electrode placement. To increase focus on an area of interest, additional scalp or subdural electrodes may be placed, including the use of cortical electrode strips. For uses in extracranial surgeries, as in carotid endarterectomies (CEAs), EEG may be employed to monitor cortex directly at risk for ischemia. In other procedures where only more general assessment of cerebral function is required (e.g., in MEP anesthetic assessment), a simpler montage may be used. In addition, EEG monitoring is often performed along with other modalities, including SEP, where electrodes are placed that may also be used as part of the EEG montage.

•

39

For CEA surgeries, the authors employ a modified eight-channel montage that takes advantage of C3' and C4' electrodes placed for SSEP recordings, with good sensitivity for quickly detecting gross changes (Figure 2.7). The eight channels represent anterior and posterior derivations with lateral and medial representations (Fp1 to P3; P3 to O1; Fp1 to C3'; C3' to O1; Fp2 to P4; P4 to O2; Fp2 to C4'; C4' to O2). A more complex montage may be used, but with the trade-off that real-time monitoring may be more difficult to follow without clear benefits of sensitivity. For MEP monitoring, only a simpler, more general setup is required, with montages utilizing existing derivations of existing SEP scalp electrodes to assess for burst suppression (Figure 2.8).

Patterns As in other NIOM modalities, EEG changes are assessed compared to the patient’s baseline and particular attention is paid to

FIGURE 2.7 A sample snapshot of an EEG montage used in carotid endarterectomy procedures employing eight channels to include frontal and posterior channels on medial and lateral aspects.

Husain 02

40

1/17/08

•

11:52 AM

Page 40

SE C T I ON I : B a s i c P r i n c iple s

FIGURE 2.8 In MEP cases, a simplified EEG montage may be employed for purposes for monitoring for presence and levels of burst suppression. This sample demonstrates a burst suppression pattern.

new asymmetries that develop during the operative procedure. Both amplitude and frequency composition may change from an intraoperative insult. As cortex is depressed, say from ischemia, lower frequencies may be observed (delta or theta), while higher frequencies (alpha or beta) are reduced or lost. Conversely, the amplitude of the EEG may initially increase as the underlying cortex becomes more synchronous under these adverse conditions. As ischemia progressively worsens, amplitudes will depress, as in conditions of suppression or cerebral silence. The sequential pattern change from normal EEG to slowing to suppression to silence reflects predictable underlying levels of cerebral hypoperfusion (8). During CEA procedures, increased slowing ipsilateral to the surgical side may be seen following carotid cross-clamping, typically with a frontal predominance. However, bilateral changes may also result from unilateral occlusion and it is helpful to ensure a stable anesthetic regimen around the time of crossclamping to help distinguish surgical- from anesthetic-induced changes. In order to ensure reasonable sensitivity for detecting changes in EEG, levels of anesthesia should ideally not induce a level that overly suppresses EEG waveforms. Particular attention may be paid to timing of delivering boluses of anesthetic agents (e.g., around time of anticipated

carotid clamping during CEA procedures) to prevent undue EEG suppression. The burst suppression pattern is characteristic of anesthetic boluses or of excessive anesthetic depth. (Figure 2.8). Burst suppression seen during MEP procedures can alert the anesthetist to excessive anesthetic and explain degraded MEPs due to inhibition at the anterior horn cell. For intracranial procedures, a state of burst suppression or silence may be chosen as a means of cerebral protection. In these cases, EEG monitoring may be helpful in titrating the depth of this suppression.

Artifact As with other NIOM modalities, achieving acceptable signal-to-noise ratios may be a challenge within the operating room. Common sources of NIOM external electrical noise include those devices, leads, and cables located near the head of the patient. Especially during periods of electrocautery, EEG monitoring is precluded. In addition, mechanical artifacts from surgeon movement of EEG leads, adjacent wires, or the patient while working near or on the skull may also contribute to EEG artifacts.

Processing Methods Continuous EEG patterns are sufficient for the experienced examiner to follow during

Husain 02

1/17/08

11:52 AM

Page 41

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

NIOM to detect changes and patterns of interest. However, several methods of EEG processing exist that may assist in interpreting the EEG for changes. These methods should not replace monitoring raw, real-time EEG in order to avoid misinterpretation of artifact that may mask changes or result in false-positive results. A common method of processing is to transform the raw EEG signal into frequency spectra to separate out the EEG frequency ranges graphically. Fourier analysis is applied to the raw EEG traces and the power of the signal at a frequency is plotted and updated over time. Other processing methods have been introduced that are meant to process EEG and analyze for the depth of anesthesia, the details of which are beyond the scope of this chapter.

Applications EEG may be used in NIOM for procedures where the cortex is directly at risk or, as discussed earlier, as a means to indirectly assess anesthetic effects. Procedures such as the resection of an arteriovenous malformation (AVM), aneurysm repair, and carotid endarterectery (CEA) are examples where vascular supply to the cortex is at risk for compromise. Often EEG is combined with other modalities such as SEPs in these procedures. When burst suppression is induced, as for cerebral protection, EEG can identify the presence of that state and its depth. The depth of burst suppression can be roughly estimated by the length of interburst suppressive period, with a typical target range of 10 to 15 seconds. Direct cortical EEG recording (electrocorticography) may be used to direct resection of epileptogenic tissue in epilepsy surgery or as part of eloquent cortex mapping techniques used to guide resection of dysfunctional brain.

Changes of Significance Once a reliable EEG baseline is obtained and an anesthetic regimen stabilized, changes in amplitude and frequency composition will

•

41

give clues to surgically induced changes affecting the cerebral cortex. As with all NIOM modalities, the timing and nature of the change as well as correlative evidence assist with interpretation of a clinically significant event. When looking for evidence of significant cerebral hypoperfusion, as during carotid endarterectomy, typical criteria indicating the need for carotid shunting are 50% loss of overall amplitude, 50% loss of alpha and beta activity, or a doubling of low-frequency activity (9).

BRAINSTEM AUDITORY EVOKED RESPONSES Background BAEPs reflect peripheral and brainstem generators within the pathway following vestibulocochlear nerve (CN VIII) activation. The first five waves of the response are resistant to anesthesia and therefore well suited to NIOM. Furthermore, the multiple generators represented allow relative localization of insults during brainstem and CN VIII pathway procedures. In addition to classic BAEP recordings, direct recording of a CN VIII nerve action potential may be performed and allows robust and rapidly acquired auditoryinduced signals.

Auditory Pathways Although exact generators may be in dispute, the auditory pathways involved in classic BAEP recordings include the distal CN VIII (wave I of the BAEP) to the cochlear nucleus (wave II; with perhaps contribution in part from cochlear nerve), superior olivary nucleus (wave III), lateral lemniscus (wave IV), and inferior colliculus (wave V). For NIOM, waveforms I to V may be reliably followed, representing a pathway from cochlear nerve to midbrain. Later waveforms generated in BAEP recordings presumably arise from thalamus and cortical radiations to auditory

Husain 02

42

1/17/08

•

11:52 AM

Page 42

SE C T I ON I : B a s i c P r i n c iple s

cortex, but these are typically absent or unreliable, especially under conditions of anesthesia. The later components may be useful as a marker of depth of anesthesia, but these uses are still investigational.

recording electrode may be placed directly on the nerve. This allows an averaged CN VIII nerve action potential to be directly recorded using the same click stimuli as described above. However, far fewer stimuli are required in recording a stable averaged response (10).

Stimulation and Recording Techniques BAEPs are elicited by delivering a click stimulus to the ear being monitored. To avoid contribution from the contralateral ear, masking white noise is introduced to the contralateral ear at approximately 60 dBnHL. Many hundreds of averaged trials are required to record reliable potentials; thus long acquisition times must be balanced against the loss of waveform resolution that occurs as the stimulation rate is unduly increased. A rate of approximately 11 Hz provides a reasonable balance. As with all signal-averaging techniques, stimulation rates should avoid divisors of 60 so as to avoid incorporation of 60-Hz noise within the average. BAEP morphology is dependent upon the polarity of the squarewave pulse delivered. A rarefaction (pulling pulse on the eardrum) or condensation (pushing pulse) may be chosen with typically an enhanced wave I from the former. Two channels are used to record BAEPs (1: ipsilateral ear – Cz; 2: contralateral ear –Cz), and others may be added as desired (e.g., ipsilateral ear–contralateral ear). Wave I is a negative potential recorded at the ipsilateral ear, while the later waves are positive potentials widely distributed across the scalp and recorded at the vertex for ease of electrode application and to minimize EMG interference. The channel from the contralateral ear is reflective of the ipsilateral channel, but without the presence of wave I and typically with better separation of waves IV and V. This information may be helpful in elucidating the different waveforms, especially for an ear with deficits that exhibits dispersed and delayed peaks. If CN VIII is at risk and exposed during surgery, as may be the case with cerebellopontine angle tumor procedures, a specialized

Interpretation As with other NIOM modalities, BAEPs are compared throughout the procedure to baseline recordings performed during early stages of the procedure. It may also be helpful to have preoperative BAEP studies to help with identification of waveforms and to provide reasonable expectation as whether signals will be present intraoperatively. A pretest expectation of signals may be particularly helpful when an intraoperative absence of all waveforms is identified at baseline. As may be reasoned, BAEPs are similarly affected by the inability to deliver an effective sound stimulus to the cochlea and by dysfunction of the cochlea. Thus the absence of wave I and subsequent waveforms is equally explained by either poor transduction of the sound pulse through the ear or dysfunction of the cochlea. The former may arise from a myriad of technical issues, including too low or absent a stimulus intensity from the sound generator, dislodgement of the tubal insert, or fluid/wax within the ear dampening the sound pulse. Thus a prior expectation of signal presence may prompt additional steps to identify and correct technical issues, while an expectation of an absence of signals can avoid extraordinary troubleshooting measures. Similarly, if all waveforms including wave I are degraded intraoperatively, technical issues must be distinguished from pathophysiologic conditions. Given the potential nonphysiologic causes, a full evaluation of stimulus and recording systems is performed. However, when abrupt loss of wave I occurs and is correlated to surgical events, it often reflects loss of cochlear function; the most common cause is compromise of the internal

Husain 02

1/17/08

11:52 AM

Page 43

CH AP T E R 2 : B a s i c N e u ro p hys iologic Intraope rative Monitoring Te chnique s

auditory artery. When this loss of all waveforms is due to surgical action, the risk of deafness is high. Persistence of wave I with loss of all later components suggests dysfunction at the level of the cochlear nerve. This pattern may reflect total nerve transection, transient conduction block, or desynchronization of the conducted auditory activity. Thus prognosis is more variable in this case and will depend on the underlying cause. Waves subsequent to wave I are generated in the brainstem, with waves III and V being the most reproducible. Changes in latencies and amplitudes of these components help to localize dysfunction of corresponding brainstem structures. Tracking the interpeak latencies of waves I to III and III to V assists in discerning anatomic areas of dysfunction. A signal delay within the cochlear nerve, for instance, may present with prolongation of waves I to III, while isolated prolongation of the III to V interpeak latency may be seen with brainstem retraction. In general, a loss of wave V amplitude greater than 50% or a 0.5-ms prolongation of its latency is considered a significant change in BAEP. However, refinement of these arbitrary levels is likely needed and may be have to be tailored to the type of surgery (11). As with other modalities of NIOM, the timing of observed changes is important in interpreting changes in relation to both surgical events and other monitored signals. For example, if changes in BAEP accompany changes in SEP, a brainstem insult is likely.

Indications Operations that involve the cerebellopontine angle are the most common ones requiring BAEP monitoring as part of a multimodality monitoring regimen. Brainstem tumor and vascular procedures, microvascular decompression, vestibular neurectomy, and other posterior fossa procedures may benefit from BAEP NIOM. Changes in BAEPs

•

43

may be seen from irreversible actions (e.g., cochlear nerve transection) or from reversible causes (such as brainstem retraction), so that timely information to the surgical team may be of assistance to spare function. Indeed, the duration of deteriorated waveforms is suggested to have a predictive correlation to clinical outcome (12). The use of direct recordings of cranial nerve VIII allows more rapid feedback and may allow for additional information when artifact would interfere with BAEP identification or when BAEPs are poorly formed but hearing is preserved.

Anesthesia and Other Factors Affecting BAEP Commonly used anesthetics including volatile agents have little or no effect on BAEP, making these waveforms relatively unique in their resilience to fluctuating anesthetics. However, BAEPs are susceptible to influences of auditory noise sources in the operating room that may not affect other NIOM modalities. Sound and electrical sources that affect the BAEPs include the presence of drilling, ultrasonic aspirators, electrocautery, intraoperative microscopes, and EMG artifact. The first three on this list are the most common offenders of interference and BAEP acquisition may need to be paused during their use. Frequently BAEPs are performed during acoustic nerve surgeries, where facial nerve EMG is also monitored and thus neuromuscular blockade is low or absent. The placement of recording electrodes at the vertex and either ear helps to minimize EMG interference, as little muscle mass exists under these electrodes. Body temperature will influence BAEP latency. In particular, with aggressive cooling conditions (e.g., < 27°C), the BAEP will be obliterated.

CONCLUSIONS Many different neurophysiologic testing modalities can be used for NIOM. Often more

Husain 02

44

1/17/08

•

11:52 AM

Page 44

SE C T I ON I : B a s i c P r i n c iple s

than one modality will be needed during a surgical procedure. It behooves the neurophysiologist to understand the surgical procedures and know which neural structures will be at risk. Based on this assessment, a combination of various modalities should be chosen for NIOM. A clear understanding of the various neurophysiologic tests in the diagnostic laboratory helps the neurophysiologist interpret them when they are used in the operating room.

6.

7.

8.

REFERENCES 1. Prass RL, Luders H. Acoustic (loudspeaker) facial electromyographic monitoring: Part 1. Evoked electromyographic activity during acoustic neuroma resection. Neurosurgery 1986;19:392–400. 2. Calancie B, Madsen P, Lebwohl N. Stimulusevoked EMG monitoring during transpedicular lumbosacral spine instrumentation. Initial clinical results. Spine 1994;19:2780–2786. 3. Nuwer MR, Dawson EG, Carlson LG, et al. Somatosensory evoked potential spinal cord monitoring reduces neurologic deficits after scoliosis surgery: results of a large multicenter survey. Electroencephalogr Clin Neurophysiol 1995;96:6–11. 4. Kothbauer K, Deletis V, Epstein FJ. Intraoperative spinal cord monitoring for intramedullary surgery: an essential adjunct. Pediatric Neurosurgery 1997;26:247–254. 5. Sala F, Niimi Y, Berenstein A, Deletis V. Neuroprotective role of neurophysiological

9. 10.

11.

12.

monitoring during endovascular procedures in the spinal cord. Ann N Y Acad Sci 2001;939: 126–136. Lyon R, Feiner J, Lieberman JA. Progressive suppression of motor evoked potentials during general anesthesia: the phenomenon of “anesthetic fade”. J Neurosurg Anesthesiol 2005; 17:13–19. MacDonald DB. Safety of intraoperative transcranial electrical stimulation motor evoked potential monitoring. J Clin Neurophysiol 2002;19:416–429. Sundt TMJ, Sharbrough FW, Piepgras DG, et al. Correlation of cerebral blood flow and electroencephalographic changes during carotid endarterectomy: with results of surgery and hemodynamics of cerebral ischemia. Mayo Clin Proc 1981;56:533–543. Nuwer MR. Intraoperative electroencephalography. J Clin Neurophysiol 1993;10:437–444. Møller AR, Jannetta PJ. Monitoring auditory functions during cranial nerve microvascular decompression operations by direct recording from the eighth nerve. J Neurosurg 1983;59: 493–499. James ML, Husain AM. Brainstem auditory evoked potential monitoring: when is change in wave V significant? Neurology 2005 22;65: 1551–1555. Nakamura M, Roser F, Dormiani M, et al. Intraoperative auditory brainstem responses in patients with cerebellopontine angle meningiomas involving the inner auditory canal: analysis of the predictive value of the responses. J Neurosurg 2005;102:637–642.

Husain 03

1/17/08

11:55 AM

3

Page 45

Remote Monitoring

Ronald G. Emerson

O

mitted across the Internet are broken into discrete packets that can follow different routes between sender and receiver. Internet data packets are even permitted to arrive out of order, to be resequenced at the destination. The influence of the Internet is so pervasive that protocols evolved to serve it govern almost all computer networks regardless of whether they actually connect to the Internet. Networked computers are identified by their IP (Internet Protocol) address. Much like a street address, an IP address indicates the location of the computer on the network. IP addresses are usually assigned by special servers that are part of the network infrastructure; the network infrastructure also implements a set of rules for routing data packets between computers based on their IP addresses. Just as the NIOM team needs to be able to detect and resolve technical problems that compromise monitoring, it is helpful for the team to be able to recognize and fix simple problems with remote monitoring. To determine the IP address of a Windows-based NIOM system, simply click Start > Programs > Accessories > Command Prompt and type “ipconfig” (Figure 3.1A). The IP address of the NIOM machine may remain constant, or it may change from day to day; within an institution, changes usually affect only the

ptimal supervision of neurophysiologic intraoperative monitoring (NIOM) and the ability to respond in a timely manner to problems and emergencies require that the clinical neurophysiologist have realtime access to NIOM data. When the neurophysiologist is not present physically in the operating room (OR), a remote data connection can provide suitable access. Indeed, many vendors of NIOM equipment also supply remote monitoring software. In hospitals with modern network infrastructures, remote monitoring within the facility can be as simple as installing these programs. Monitoring from more distant sites or from sites with suboptimal or more complex data connections can be more complicated. Nonetheless, in most cases a suitable solution is available.

BACKGROUND Although it is not the intent here is not to offer a tutorial on the computer networking or the Internet, a few necessary basics are covered. Today’s Internet is a direct descendent of an initiative from the Cold War era to develop a robust communication network that would survive wartime damage. In contrast to traditional telephony, where communication is continuous between two points, data trans45

Husain 03

46

1/17/08

•

11:55 AM

Page 46

SE C T I ON I : B a s i c P r i n c iple s

A

B

FIGURE 3.1 To determine the IP address of a Windows computer, (A) first open a Command Prompt Window by clicking Start à Programs –> Accessories –> Command Prompt. (B) in the Command Prompt Window, type “ipconfig.” The system will respond with information that includes the IP address of your machine and also the IP address of your default gateway.

rightmost digits (e.g., 21 in Figure 3.1B). An IP address of 0.0.0.0 or, by convention, 169.254.x.x1 is invalid and usually indicate that your computer was unable to obtain a proper address. An easy way to verify connec1

In this notation, x is an integer from 0 to 255.

tion to the network and that the network is working is to send a test message (or “ping”) to some other machine on the network. If the message is received, the target machine will respond with the time it took for the test message to be received. If the message is not received, the ping will fail.

Husain 03

1/17/08

11:55 AM

Page 47

CH A P T ER 3: R e mote Monitoring

•

47

FIGURE 3.4 Network icon in the System Tray, usually at the lower-right-hand corner of a Windows computer. FIGURE 3.2 To ping a machine, open a Command Prompt Window (see Figure 3.1), and type “ping,” followed by the IP address of the target system. In this example, the target is 10.2.2.16. The ping is successful, with times between 1 and 2 milliseconds.

The “default gateway,” the device that typically connects the computers on a floor or building with a larger network or to the Internet, is often a good choice for pinging. “Ipconfig” also displays a machine’s address (Figure 3.1B). In Figure 3.2, the command “ping 10.2.2.16” sends a series of four test messages to the default gateway; in each case the gateway machine responds promptly. In contrast, in Figure 3.3, the connection has been lost and the ping fails. On Window machines, the network connection can be partially tested by clicking on the network icon in the system tray, usually at the lower right of the screen (Figure 3.4). A bad network connection may return a message indicating “limited connection” or “network cable unplugged.” A defective network cable can be physically plugged in but functionally disconnected; loss of resilience with use of the plastic tab on the network connector is a common cause of an unreliable or

FIGURE 3.3 Example of a failed ping, returning the message “request timed out.”

intermittent connection on portable equipment (Figure 3.5). Local networks are generally designed to permit data to flow freely between all connected devices. At points where a local network joins a larger network or the Internet, routers and firewalls may selectively limit the flow of data. Some types of data may be permitted to reach only specific destinations, while others may be excluded entirely. Generally, institutional firewalls have a “deny all” policy for inbound traffic, with exceptions provided for special data types. The firewall illustrated in Figure 3.6 block all inbound traffic but web and mail data; designated servers are the only permitted destinations. Filtering by data type is easily accomplished because, in addition to the data “payload” and the IP addresses of sending and receiving computers, data packets also specify source and destination “port numbers,” which effectively specify the type of data car-

FIGURE 3.5 Ethernet cable connector. The arrow points to a plastic tab that commonly loses resilience with age, causing an unreliable electrical connection.

Husain 03

48

1/17/08

•

11:55 AM

Page 48

SE C T I ON I : B a s i c P r i n c iple s

FIGURE 3.6 Institutional firewall blocks all inbound traffic except for packets bound for the mail and web server. Packets that are inbound but are responding to messages sent by inside computers are also permitted.

ried in the payload (Figure 3.7). In Figure 3.6, the firewall is configured to permit inbound traffic only to the web server on Port 80 and the mail server on Port 25. By convention, IP numbers in the 192.168.x.x and 10.x.x.x ranges are nonroutable on the Internet and reserved for use on private networks. Commonly, institutions make use of nonroutable IP numbers internally; the network address translation (NAT) protocol is used by the firewall to enable internal machines to communicate with external hosts. This arrangement makes internal machines invisible to the Internet unless communication is initiated by the internal system or special firewall provisions are made to “map” specific internal machines to additional, routable IP addresses.

FIGURE 3.7 Typical data packet consisting of a “data payload” plus the source and destination IP addresses and ports.

REMOTE CONNECTION TECHNIQUES Various options exist for remote data connections. The best choice will depend on the type and quality of the remote Internet connection and the services and policies of the hospital.

Virtual Private Networks One approach entails the use of a virtual private network (VPN). A VPN is effectively a secure tunnel through the Internet connecting the hospital’s internal network and a remote site. A VPN makes the computer (or an entire remote site) appear as if it were within the hospital, behind its firewall (Figure 3.8). If the hospital uses private (192.168.x.x or 10.x.x.x) IP addresses, these can be directly accessible on a VPN. Technically, virtual private networking is accomplished by wrapping the internal network packets (including both internal IP addresses and the data payload) inside another packet whose IP addresses specify the ends of the tunnel (Figure 3.9). In this arrangement, the remote computer has

Husain 03

1/17/08

11:55 AM

Page 49

CH A P T ER 3: R e mote Monitoring

•

49

FIGURE 3.8 A VPN creates a secure tunnel through the Internet, allowing an external computer to function as if it were behind the institutional firewall.

two IP addresses: a normal Internet IP address and an IP address that is actually part of the hospital network. Generally, the internal network packets are encrypted before they are wrapped inside the VPN packets and are decrypted as they emerge from the tunnel. Security concerns may cause some institutions to not offer or restrict remote VPN connections. Since a VPN logically moves the remote computer to within the institutional firewall, it also exposes the internal network to viruses and other potentially dangerous software that may be present on remote machines. Although desktop computers within an institution may be actively managed, kept up to date with antivirus software and security patches, and restricted to running properly vetted software, a similar level of control of home computers is likely to be more difficult.

VPN tunnels can be intolerant of suboptimal network connection. Communication glitches, common on the Internet, may interrupt VPN traffic, requiring the user to reestablish connection. VPN may not work in some settings (e.g., on some hotel guest networks).

Remote Application Servers An alternative approach to remote access is to run the applications to be accessed remotely on special remote application servers. When remote users connect, their computers become the virtual keyboards and monitors of a remote server on which the application actually runs. Only the image to be displayed plus keyboard and mouse clicks are transmitted between the server and the

FIGURE 3.9 VPN data packet. An internal data packet is encrypted and becomes the data payload carried by the VPN “wrapper.”

Husain 03

50

1/17/08

•

11:55 AM

Page 50

SE C T I ON I : B a s i c P r i n c iple s

FIGURE 3.10 Only virtual screen images plus keyboard and mouse clicks travel beyond the institutional firewall in this remote application server implementation. Although it appears to the user that applications are running on the remote computer, they are really running on the remote application server.

remote user (Figure 3.10). If desired, remote users can be permitted to view documents but prohibited from transferring files. In contrast to VPN, which can expose internal networks to viruses and other security threats posed by unmanaged remote computers, the remote application server solution can fully shield the internal network. Additionally, remote application systems can require less bandwidth and can offer better performance under suboptimal Internet conditions. Conditions that would cause a VPN to disconnect may cause remote application systems to simply stutter or become momentarily unresponsive. On the other hand, given sufficient bandwidth and stable Internet connectivity, some software may perform better over a VPN. Although remote desktop systems require special client software to be present on the remote system, this client software often requires minimal or no user configuration. Microsoft’s Windows Servers 2000 and 2003 support remote application services. Citrix Systems offers a well-known remote application solution based on Microsoft’s

Windows Terminal Services but may provide better security and performance (1,2). Citrix also provides a web-based interface that also checks for the client software on the remote machine; if it is not there, the user is directed through a very simple one-time installation process.

REMOTE NIOM REVIEW SOLUTIONS In addition to establishing a remote connection with the NIOM machine, a technique must exist to transmit the data being collected for remote NIOM review. Several techniques are available for such review.

Vendor-Specific Remote Display Software Vendors of NIOM systems commonly provide companion programs for remote display of NIOM data in real time over a network connection. These programs typically emulate the display and review functions of

Husain 03

1/17/08

11:55 AM

Page 51

CH A P T ER 3: R e mote Monitoring

the NIOM machine. They permit the remote user to view ongoing recordings in real time as well as to review prior data from earlier in a case. Since the actual physiologic waveform data are sent to the remote site, the user is typically able to manipulate the remote display (e.g., to change gains and filters independently of the technologist in the OR).

In-Hospital Use of Vendor-Specific Remote Display Software Vendor-specific remote software generally functions seamlessly on internal hospital networks. There are, however, a few important technical points of which the user should be aware. If the NIOM or remote machines have “personal firewalls,” it will probably be necessary to open the port used by the remote software. Personal firewalls are enabled by default in Windows XP service pack 2. To use the remote software, the user typically needs to know the IP address or name of the NIOM machine. Depending upon institutional policy and whether the NIOM machines is used in more than one location within the hospital, the NIOM machine will be assigned either a fixed IP address or will receive a “dynamic” address that may change from day to day. In the latter case, the technician in the OR will need to make the remote user aware of the machine’s IP address. The remote software may also permit the user to specify the NIOM machine by name. A machine may be identified by an official “fully qualified” name (e.g., NIOM2.neuro.columbia.edu), by a simple shorter name (NIOM2), or by both. Depending upon how the NIOM machine is configured and the network naming protocols in use, the remote user may have better success using the IP number than the name.

Remote Use of Vendor-Specific Remote Display Software Special provisions are likely to be necessary for using vendor-specific remote software

•

51

from beyond the hospital firewall. Typically, the firewall will prevent software running on an outside computer from initiating communication with the NIOM machine in the OR. A VPN is commonly used to access to a hospital’s internal network from outside its firewall. Vendor-specific remote client running over a VPN can be the ideal remote monitoring solution, providing the remote monitoring station with full functionality—the same as would be available from within the hospital. The requirements are (a) support of the VPN by the hospital, (b) installation of both VPN “client” and vendor-specific remote software on the remote computer, and (c) a stable, highquality Internet connection to the remote site. Use of a remote application server can be a reasonable alternative if it is compatible with the vendor-specific remote application. Because one remote system will effectively be running on top of another, remote monitoring may seem somewhat sluggishly responsive to key and mouse clicks. Nonetheless, a remote application server implementation has the advantages of (a) only a single license for the vendor—supplied remote software may be necessary, (b) less software—and no NIOM specific software—needs to be installed on the remote computer so that virtually any Internet-connected computer can be a remote monitoring station, and (c) it may function more reliably than a VPN if the Internet connection is slow or otherwise suboptimal.

Screen Capture Remote Control Software Generic screen-capture software is an alternative to a vendor-specific remote program. Typically, the “server” version of the software is installed on the NIOM machines and the “client” version on the remote machine. Rather than transmitting waveform data to the remote site, screen-capture software simply transmits an image of the NIOM machine’s screen. An obvious limitation to this is that the remote station can see only

Husain 03

52

1/17/08

•

11:55 AM

Page 52

SE C T I ON I : B a s i c P r i n c iple s

what is happening in the OR in real time; the user cannot replay past events or change settings independently. The corresponding advantage is that the remote station displays exactly what is happening in the OR without requiring the user to choose settings or make adjustments. Assuming that the remote user has simultaneous voice or “instant messaging” communication with the technician in the OR, it is simple to say “show me.” A wide selection of screen-capture remote control software is commercially available. One of the more ubiquitous is virtual network computing (VNC). Several variants of VNC are available free under the General Public License.2 VNC is platformindependent, meaning that you can, for example, run a Windows version on the NIOM machine in the OR and monitor remotely using an Apple or Linux machine. Although the basic version of VNC does not support encryption, remote communication can be secured by simply running VNC (or other screen-capture remote control software) through a secure VPN tunnel. Rather than using a VPN, the screen-capture client software may be installed on a remote application server. Again, performance here may be sluggish because two layers of remote software are in use. Yet if a VPN connection is not available or suitable and if vendor-specific software is either not available or not compatible with the remote application server, this arrangement may be acceptable. pcAnywhere and other similar commercial products support secure, encrypted remote connection. Proper configuration of this software is critical, since it is possible to configure it in a nonsecure manner. Also, it is likely that specific ports will need to be opened in the hospital’s firewall. Users should

2 Supported by the Free Software Foundation (www.fsf.org), the General Public License grants users access to run, to study how a program works, and to modify it and distribute it; it and also provides that these right will be preserved in derivative works.

therefore consult with appropriate authorities at their institutions. An important caution when using screencapture software for remote monitoring is to avoid inadvertently granting control permission to the remote site. By design, VNC and similar software packages allow the remote user to control the host system—i.e., a user at home could control the NIOM machine in the OR. If that is not intended, this option should be disabled.

Reverse-Connection Remote Software A popular variant of the screen-capture remote control software makes it easy to connect from a remote computer to a target computer behind a hospital firewall. Programs such as GoToMyPC circumvent firewall restrictions by reversing the direction of communication initiation (3). As discussed above, most firewalls are generally configured to permit only very specific inbound connections and block all others. However, they are more permissive with outbound connections, and almost all institutional firewalls permit outbound web traffic (port 80 for unencrypted and 443 for encrypted). Once an outbound conversation has been initiated, associated inbound traffic is permitted (otherwise one could not browse the web). In the reverse-connection scenario, special software on the target computer (e.g., the NIOM machine in the OR), in addition to performing screen capture and remote server functions, also establishes and maintains a continuous connection with an outside thirdparty server by sending it periodic “keep alive” packets (Figure 3.11). When a remote user wants to contact the target computer, it instead contacts the third-party server; the remote user is then able to communicate with the target machine by effectively joining a conversation already in progress, initiated from within the firewall! Reverse-connection remote systems can offer elaborate security, including remote user authentication, end-to-end data encryption so

Husain 03

1/17/08

11:55 AM

Page 53

CH A P T ER 3: R e mote Monitoring

•

53

FIGURE 3.11 Reverse-connection remote system software on the target computer (e.g., the NIOM machine) sends “keep alive” packets to a third-party server, which waits for the remote system to connect.

that the third party cannot decipher the data being transferred between the target and remote machines, and provisions to prevent target systems from becoming unknowing hosts to reverse-connection remote software. Nonetheless, this technology does provide a mechanism for potentially violating the intent of institutional firewall and security policies; in any case, it transfers responsibility for the security of data protected by the Health Insurance Portability and Accountability Act (HIPAA) to a third party. Accordingly, some institutions may block access to these thirdparty servers. For these reasons, it is suggested that users consult with their institutions prior installing this type of software for remote NIOM monitoring.

Image Capture/Web Server A “solution of last resort,” one that will work in almost any environment, is to install software on the NIOM machine that periodically takes a picture of the screen display (or a portion of it) and automatically sends it to a web server. The web server can be located

behind the institutional firewall or, if external access to an internal web server is not easily arranged, on the Internet using an independent Internet service provider. Inexpensive programs, such as Snagit, can be set to automatically capture screen images at preset intervals (e.g., twice a minute) and output them, for example, using the file-transfer protocol (FTP). A very simple script causes the image displayed by the browser to be refreshed at corresponding intervals. Configuration of the remote server as both a web server to display the images and an FTP server to receive them, while beyond the scope of this chapter, is simple and is well within the capabilities of many computer novices. A disadvantages of this technique is that screen images are captured periodically (e.g., once of several times a minute, not continuously). Further, unless secure, encrypted protocols are used for both the web connection and transfer of the images from the NIOM machine to the web server (FTP is not secure), it is critical that no patient identifiers (e.g., the patient’s name, unit number, etc.) be inadvertently captured in the image.

Husain 03

54

1/17/08

•

11:55 AM

Page 54

SE C T I ON I : B a s i c P r i n c iple s

An advantage of this technique is that it will work well with just about any remote computer with an Internet connection. No special software needs to be present on the remote system; a web browser is all that is necessary. Web connections are, by design, not continuous; when you browse the web, each mouse click typically initiates an independent conversation with the server that ceases once the page is transmitted. Accordingly, this type of connection is suitable when Internet connectivity is slow or unreliable, and it works well with handheld wireless devices.

CONCLUSIONS In today’s NIOM environment, remote review of data by the neurophysiologist is common. The NIOM team must be knowledgeable not only about monitoring techniques but also about Internet connectivity to ensure that data are adequately transmitted from the acquisition system to the remote

review system. There are several different ways in which connection to the host computer can be established, both from within the hospital and from the outside. There are many different methods that allow data from the acquisition system to be reviewed at a remote site. Each has advantages and disadvantages and each NIOM team must pick a system that works well for them.

REFERENCES 1. Montoro M. Microsoft RDP man in the middle vulnerability. Available at: http://www .securiteam.com/windowsntfocus/5EP010KG 0G.html. Accessed November 20, 2006. 2. Frequently Asked Questions—Citrix ICA and man-in-the-middle attacks. Available at: http://support.citrix.com/article/CTX101737. Accessed November 20, 2006. 3. Citrix GoToMyPC security white paper. Available at: https://www.gotomypc.com/ downloads/pdf/m/GoToMyPC_Security_ White_Paper.pdf. Accessed November 20, 2006.

Husain 04

1/17/08

11:55 AM

Page 55

4

Anesthetic Considerations

Michael L. James

T

he practice of anesthesia has historically relied on the induction of a reversible state of amnesia, analgesia, and motionlessness. With the improvement of medical technology, advancement of knowledge, and practice of evidence-based medicine, modern anesthesiology comprises a great deal more. It has become the role of the anesthesiologist during surgical, obstetrical, and diagnostic procedures to provide anesthesia, optimize procedural conditions, maintain homeostasis, and, should it be necessary, manage cardiopulmonary resuscitation. Additionally, anesthesiology has found itself branching out into chronic and acute pain treatment as well as the intensive care unit. Obviously there has been an expansion of expectations for the practice of anesthesia over the last few decades; however, ultimately, anesthesiology is the practice of manipulating a patient’s neurologic system and physiology to effect some beneficial end.

consists of four basic stages: premedication, induction, maintenance, and emergence. Prior to entering the operating suite, “premedications” may be administered to prepare the patient for the perioperative period. Usually this takes the form of mild sedation for anxiolysis, analgesics for preprocedural pain, antihypertensives, antiemetics for patients with a high likelihood of postoperative nausea and vomiting, antisialagogues to facilitate intubation, etc. In the operating room the historic principles of anesthesia are still the foundation of practice, and analgesia (i.e., painlessness), amnesia (i.e., memory loss), motionlessness, and hemodynamic stability can be obtained and maintained by a variety of means. Commonly, general anesthesia is induced through the administration of a large bolus dose of an intravenous sedative-hypnotic (e.g., propofol). A dose of intravenous opioid (e.g., fentanyl) and a paralytic agent (e.g., vecuronium) may be given at this time as well to facilitate endotracheal intubation. After induction, anesthesia maintenance usually consists of some amount of inhaled volatile anesthetic agent (e.g., isoflurane) in a mix of oxygen and either air or nitrous oxide and some dose of intravenous opioid. The amount of volatile agent is quantified in terms of mean alveolar concentration (MAC). MAC is expressed as a percentage of inhaled gas and is defined as the

PRINCIPLES OF ANESTHESIA There are four basic types of “anesthesia”: general anesthesia, regional anesthesia, local anesthesia, and sedation. For the purposes of neurophysiologic intraoperative monitoring (NIOM), general anesthesia (the creation of reversible coma) is nearly always required and 55

Husain 04

56

1/17/08

•

11:55 AM

Page 56

SE C T I ON I : B a s i c P r i n c iple s

alveolar partial pressure of a gas at which 50% of patients will not move with a 1-cm abdominal surgical incision. However, in practice the necessary amount of volatile agent is determined by effect. It is during anesthesia maintenance that NIOM occurs (as does the surgical procedure). After the procedure is finished, the expectation is that the anesthetic coma will be completely reversible, and the patient must emerge from anesthesia without experiencing lasting effects from the agents. Emergence is usually accomplished by reversing any residual neuromuscular blockade and allowing the patient to eliminate volatile agent via breathing. Volatile anesthetic agents are minimally metabolized and largely removed from the body in the same manner they were introduced: ventilation. In terms of NIOM, special considerations for general anesthesia are discussed later; however, it is important for neurophysiologists and technologists to have a clear expectation of the step-by-step nature whereby anesthetic and surgical procedures are undertaken, and it is important to remember that the operating room is generally a highly active environment with people, monitors, equipment, and electrical cords all moving about at once. Any change in the NIOM may be due to many factors, not the least of which is the surgical procedure, and every attempt should be made to regain fading or lost waveforms, as permanent loss may indicate postoperative impairment (1). Therefore the entire process becomes most efficient when each individual in the room understands all the steps, including those of every other individual, required to prepare for, perform, and enable emergence from a procedure in an environment of open communication and respect for each other’s responsibilities.

NONPHARMACOLOGIC FACTORS: ANESTHETIC CONSIDERATIONS Physiologic function of the human body plays a major role in neuronal functioning; it

is arguably the most important factor, and a great deal of human physiology is influenced by actions of the anesthesiologist. The manner in which these physiologic functions are manipulated often directly determines measurable neurophysiologic function. Further, it is reasonable to assume that physiologic function determines, in large part, the survivability of nerves and their supporting structures.

Temperature It is well established that temperature plays a significant role in nerve function, especially in the axon. Changes of a fraction of a degree can drastically alter latencies and amplitudes of neuronal potentials with cortical structures being more affected than peripheral nerves (2). Relative hypothermia produces changes that invariably present as slowed latencies from slower nerve conduction. In addition there are predictable, characteristic effects of profound hypothermia that, at least initially, begin with slowing to a delta frequency (3). The opposite is true with relative hyperthermia for both evoked potentials (EPs) and electroencephalograms (EEGs). It is important to note that regional temperature changes are invariably difficult to predict, for a variety of reasons. General anesthesia causes an overall cooling effect in the body core due to peripheral vasodilatation, which is usually opposed by active surface warming and warmed intravenous fluids. Additionally, cold and/or warm irrigants are nearly always applied to the surgical field. As a result, the extremities, brain, and spinal cord are being heated or cooled depending on where they lie in relation to warmed air blankets, intravenous fluid lines, the surgical field, etc. Therefore, unless it is individually measured, the actual temperature of a given region is impossible to know, but the potential effects should be kept in mind during the course of monitoring. It is very common for patients to experience a decrease in core body temperature for the first 15 minutes after anesthetic

Husain 04

1/17/08

11:55 AM

Page 57

CH A P T ER 4: A ne s the tic Cons id e rations

induction. With active warming during the administration of most anesthetic agents— unless the surgical procedure requires an alternative strategy—the patient’s temperature will then be kept greater than 36°C by the anesthesiologist.

•

57

tionally), patient positioning, tourniquets, vasospasm, vascular ligation, etc. Anecdotally, some have reported discovering incidental ulnar nerve ischemia secondary to compression during routine monitoring for spinal fusion. When the compression was released, the nerve potentials returned to normal.

Blood Flow Logic dictates that ischemic nerves do not function normally; therefore measurable neural potentials would become abnormal. In fact it has been demonstrated that somatosensory evoked potentials (SEPs) can be lost when cerebral blood flow falls below 15 mL/min/100 g (2). This can be assumed to be true for the spinal cord and peripheral nerves as well. Unfortunately, it is difficult to actually measure blood flow to any given structure, so systemic blood pressure is often used as a surrogate. Furthermore, systemic blood flow does not necessarily dictate regional blood flow, especially in the brain, which makes it even more difficult to predict. Monitors are becoming available that purport to quantify regional blood flow (e.g., cerebral oximetry, microdialysis), but a discussion of these is beyond the scope of this chapter. Essentially then, there are two main considerations for the neurophysiologist: systemic hypotension and decreased regional blood flow. When profound, systemic hypotension results in globally reduced blood flow, which translates into tissue ischemia of varying degrees based largely on autoregulation. For example, during spinal surgery, controlled deliberate hypotension is often requested of the anesthesiologist so as to assist in controlling blood loss; however, surgical traction and hypotension can aggravate each other with deleterious effects to the patient, and NIOM can assist in determining the acceptable limit of systemic hypotension (4). There are many examples of causes of decreased regional blood flow, and almost all are due to some interruption in blood supply either due to compression from surgical instruments (intentionally or uninten-

Ventilation Optimal neural functioning depends on maintenance of a homeostatic extracellular environment. Hypo- or hypercapnea can alter cellular metabolism by changing the acid-base status of the individual. In general individuals tolerate relatively profound acid-base derangements, especially upward trends in pH. Unless the pH of a patient drops below 7.2, neuronal mechanisms are maintained. Additionally, there is a suggestion that extremes in hypocarbia (< 20 mmHg partial pressure) can alter SEP monitoring (5). Alternatively, profound hypoxia is poorly tolerated, especially in the surgical setting of ongoing blood loss and potential hypotension.

Hematology Like hypoxia, profound anemia can contribute to neural dysfunction. Normally, anemia is well tolerated to levels of hemoglobin less than 7 g/dL. However, in the surgical setting of possible large volume blood loss, hypotension, and hypoxia, it is generally accepted that hemoglobin levels should be kept above 8 g/dL and may require optimizing at 10 g/dL. At approximately 10 g/dL of hemoglobin, oxygen delivery appears to be maximized and transfusion above this threshold does not appear to improve augmentation. There are animal data that support this supposition in SEP monitoring (6).

Intracranial Pressure Increase in intracranial pressure is a relatively well documented cause of shifts in cor-

Husain 04

1/17/08

58

•

11:55 AM

Page 58

SE C T I ON I : B a s i c P r i n c iple s

tical responses of EPs and prolongation of motor evoked potentials (MEPs), presumably due to compression of cortical structures. There is a pressure-related increase in latency and decrease in amplitude of cortical SEPs and as intracranial pressure becomes pathologic, uncal herniation occurs with subsequent loss of subcortical SEP responses and brainstem auditory evoked potentials (BAEPs) (7). Alleviation of this pressure can return EPs to normal.

Other Factors Neuronal function depends on maintenance of a homeostatic intra- and extracellular environment determined by potassium, calcium, and sodium concentrations. It is logical to assume that alteration in these concentrations would result in dysfunction and possible changes in measurable neuronal potentials. The concentration of these ions is largely in the control of the anesthesiologist, and maintenance within ranges of normal values is necessary. In addition, profound hyperor hypoglycemia should be avoided, as either extreme can result in cellular dysfunction; although there is no evidence that they result in intraoperative changes in NIOM, there are data to suggest that both can lead to poor outcomes (8).

EFFECTS OF SPECIFIC ANESTHETIC AGENTS In general the anesthesiologist and neurophysiologist are constantly at odds in that nearly all anesthetic agents, given in high enough doses, cause depression of NIOM potentials. However, with open communication and mutual understanding of each other’s activities, NIOM can be successful with nearly any anesthetic technique. The crucial concept is that any change in either anesthetic or NIOM must be communicated to the team, so that every person in the operating room is acting under appropriate assumptions.

Inhalation Agents Despite being the oldest form of anesthesia, the exact mechanism of action of inhalation agents remains unclear. Inhalation anesthetics consist of two basic gases available in the United States: halogenated agents (halothane, isoflurane, sevoflurane, desflurane) and nitrous oxide. Doses of gas are given as percentage of inhaled mixture, and effective doses are expressed as some amount of MAC. As discussed before, one MAC of an agent is sufficient to prevent 50% of patients from moving to the stimulation of surgical incision (Tables 4.1 and 4.2).

TABLE 4.1 Effects of Inhaled Agents on Evoked Potentials BAEP Agents

SEP

MEP

Latency

Amplitude

Latency

Amplitude

Latency

Amplitude

Desflurane

Inc

0

Inc

Dec

Inc

Dec

Enflurane

Inc

0

Inc

Inc

Inc

Dec

Halothane

Inc

0

Inc

Dec

Inc

Dec

Isoflurane

Inc

0

Inc

Dec

Inc

Dec

Sevoflurane

Inc

0

Inc

Dec

Inc

Dec

0

Dec

0

Dec

Inc

Dec

Nitrous oxide

Inc = increased; Dec = decreased; 0 = no change.

Husain 04

1/17/08

11:55 AM

Page 59

CH A P T ER 4: A ne s the tic Cons id e rations

•

59

TABLE 4.2 Effects of Anesthetics Agents on Electroencephalogram INCREASED FREQUENCY

SUPPRESSED

Barbiturate (low dose)

Barbiturates (high dose)

Benzodiazepine

Propofol (high dose)

Etomidate

Benzodiazepine (high dose)

Propofol Ketamine Halogenated agents (< 1 MAC) INCREASED AMPLITUDE

ELECTROCEREBRAL SILENCE

Barbiturate (moderate dose)

Barbiturates

Etomidate

Propofol

Opioid

Etomidate

Halogenated agents

Halogenated agents

(1–2 MAC)

(> 2 MAC except halothane)

Inc = increased; Dec = decreased; 0 = no change.

Halogenated Agents The halogenated agents consist of the historic agent halothane, which is still used in most countries outside the United States, and the modern agents consisting of isoflurane, sevoflurane, and desflurane. Each has its own MAC, onset and offset times, and metabolism based on the inherent properties of the gas. Their use results in a dose-related decrease in amplitude and slowing of latency of SEPs, with the least effect seen in peripheral and subcortical responses (2). BAEPs are minimally affected by halogenated anesthetics at usual doses but can be ablated at high doses. MEPs are enormously affected by the use of halogenated agents and can be entirely ablated even with doses of 0.5 MAC. It appears that this effect occurs proximal to the anterior horn cell due to evidence that waves recorded distal to the anterior horn cell and proximal to the neuromuscular junction remain recordable even at high doses of anesthetic (9). MEP monitoring may also occur through spinal or epidural stimulation with minimal effect on recorded responses; how-

ever, cord stimulation results in stimulation of the sensory and motor pathways, and halogenated gases preferentially block the motor responses (10). Therefore it is important to remember that NIOM utilizing spinal cord stimulation may not reliably monitor motor function in the presence of halogenated gases. For this and reasons mentioned above—namely, easy ablation when MEP monitoring is essential—halogenated gases should usually not be part of the anesthetic regimen when using this modality. The EEG is affected but usually without hindrance to monitoring. All halogenated anesthetics produce a frontal shift of the rhythm predominance when used at induction doses (two to three times MAC doses). The gases then produce a dose-dependent reduction in frequency and amplitude. It is important to note that both isoflurane and desflurane can produce burst suppression and electrocerebral silence at clinical doses. For practical purposes, however, all halogenated agents can be used for maintenance anesthesia when NIOM requires EEG monitoring.

Husain 04

1/17/08

•

60

11:55 AM

Page 60

SE C T I ON I : B a s i c P r i n c iple s

Nitrous Oxide Nitrous oxide is similar to halogenated anesthetic agents and causes a dose-related decrease in amplitude and prolongation of latency of cortical SEPs and ablation of MEPs. This effect seems somewhat limited in subcortical and peripheral potentials of the SEPs. At equipotent doses to halogenated agents, nitrous oxide may, in fact, cause greater EP depression (2). Additionally, nitrous oxide has somewhat indeterminate effects on the EEG that is highly dependent on other agents and doses being used simultaneously. The effects on the EEG are not wholly predictable, but generally, there is frontally dominant high-frequency activity and posterior slowing. Despite this, a frequent anesthetic technique used during NIOM is a “nitrous-narcotic” technique. The modern version of this technique consists of a high-dose remifentanil infusion (0.2 to 0.5 µg/kg/min) with 60% to 70% inhaled fraction of nitrous oxide. A high, but constant, amount of nitrous oxide is delivered with varying amounts of remifentanil based on surgical stimulation. As long as the percentage of inhaled nitrous oxide is held constant, this practice generally allows recordable responses for most NIOM except transcranial MEPs, although even then 50% to 60% nitrous oxide may be used. The benefit of using nitrous oxide is that brain concentra-

tions vary rapidly with inhaled concentrations, so that if NIOM is problematic and needs maximizing intraoperatively, discontinuance of nitrous oxide will quickly result in the its elimination from the brain and body.

Intravenous Agents Intravenous anesthetic agents are generally used to induce anesthesia and afterwards to supplement inhalation maintenance anesthesia. Most modern anesthetic techniques consist of a variety of agents, intravenous and inhaled; nearly always an intravenous opioid is administered to augment other agents for either tracheal intubation at induction or intense surgical stimulation exceeding a stable maintenance anesthesia. If halogenated agents are contraindicated or NIOM becomes problematic with their use, a complete anesthetic can consist of intravenous drugs, or total intravenous anesthesia (TIVA). TIVA exists in many forms. The most common regimen is based on continuous propofol infusion and supplementation with intravenous opioid. However, all manner of TIVAs have been described, including the use of ketamine, barbiturate, midazolam, dexmedetomidine, etc., with drug selection depending on utilizing specific attributes of an agent to effect a specific outcome (Tables 4.2 and 4.3).

TABLE 4.3 Effects of Intravenous Agents on Evoked Potentials BAEP Agents

SEP

MEP

Latency

Amplitude

Latency

Amplitude

Latency

Amplitude

Low dose

0

0

0

0

Inc

Dec

High dose

Inc

Dec

Inc

Dec

Inc

Dec

Benzodiazepine

0

0

Inc

Dec

Inc

Dec

Barbiturate

Opioid

0

0

Inc

Dec

0

0

Etomidate

0

0

Inc

Inc

0

0

Propofol

Inc

0

Inc

Dec

Inc

Dec

Ketamine

Inc

0

Inc

Inc

0

0

Inc = increased; Dec = decreased; 0 = no change.

Husain 04

1/17/08

11:55 AM

Page 61

CH A P T ER 4: A ne s the tic Cons id e rations

Barbiturates Some of the oldest intravenous anesthetics include barbiturates (e.g., thiopental, pentobarbital, phenobarbital, methohexital). These drugs exist in alkaline salt solution and exert their mechanism of action at the GABAA receptor. Of these, thiopental remains in common use, in certain surgical cases, as an induction agent and as a means of achieving neuroprotection through “burst suppression.” Additionally, methohexital is frequently used to facilitate electroconvulsive therapy (ECT). However, much like halogenated agents, barbiturates will produce EEG slowing and, at higher doses, burst suppression and electrocerebral silence. There appears to be little class effect of barbiturates on SEPs, with each agent producing somewhat different results. Thiopental produces transient decreases in amplitude and increases in latency with bolus dosing for induction, but phenobarbital produces little effect until doses causing cardiovascular collapse are reached (11). As with inhaled agents, SEP cortical potentials seemed to be most affected, with relative sparing of subcortical and peripheral responses. In contrast, whether with low-dose continuous infusion or single-bolus dosing, MEP responses can be entirely abolished with the use of barbiturates. Any anesthetic given for a surgical procedure requiring MEP monitoring should exclude the use of barbiturates in any form unless their use (i.e., neuroprotection) supersedes the benefit from MEP monitoring.

Benzodiazepines Midazolam is a common intravenous benzodiazepine used in preoperative areas prior to transfer to the operating suite. Benzodiazepines also have their site of action at the GABA receptor and have the desirable effects of amnesia, sedation, and anxiolysis. In general, single one-time doses of midazolam given prior to induction have little effect on NIOM during critical portions of the procedure. However, induction doses of midazolam (0.2 mg/kg) or continuous infusions of the

•

61

drug can slow SEP latencies and decrease amplitudes (12). Furthermore, as with most other anesthetics, even small doses of benzodiazepines (1 to 2 mg) can lead to a marked reduction in MEP responses. However, owing to relatively rapid metabolism of single adminstration, if small doses of midazolam are given preoperatively, their effects on NIOM are usually minimal. Of note, benzodiazepines are anticonvulsants and will all produce slowing of the EEG into the theta range; however, at small doses they create betarhythm predominance in frontal leads, which is also seen with chronic oral administration.

Propofol Propofol remains one of the most common agents used for the induction of anesthesia and is the most common agent used for maintenance anesthesia during TIVA. It is packaged in a lipid-soluble solution and its site of action is also at the GABA receptor. Owing to rapid redistribution after dosing, propofol is easily titratable to the desired effect, which makes it very useful for TIVA techniques. Induction doses of propofol (2 to 5 mg/kg) cause amplitude depression of EEG, SEP, and MEP responses, as does high-dose continuous infusion (80 to 100 µg/kg/min). However, there is generally rapid recovery after termination if long infusion times (>8 hours) are avoided (13). In recording SEPs or MEPs from the epidural space, there seems to be limited effect of the drug on the EPs; this seems to hold true for recordings from the scalp or peripheral muscle as well (14). Propofol is also notable as an agent that can produce burst suppression and electrocerebral silence on the EEG. Despite profound EEG suppression at high dose, propofol retains its relatively quick termination, allowing for an awake, alert, and neurologically testable patient at the end of a surgical procedure.

Opioids Intravenous opioids represent a critical mainstay in the practice of modern “balanced”

Husain 04

62

1/17/08

•

11:55 AM

Page 62

SE C T I ON I : B a s i c P r i n c iple s

anesthesia to control perioperative pain. Nearly all general anesthetics will have some form and dose of intravenous opioid as a central component. Intravenous opioids in current use during the perioperative period include morphine, hydromorphone, fentanyl, alfentanil, sufentanil, and remifentanil; they are administered for various indications and at a wide variation in dosing regimens. All intravenous opioids have almost no effect on NIOM even at very high doses, making them of essential importance during anesthesia for procedures requiring NIOM. Even when given in the epidural or intrathecal space, they have minimal effect on EPs (2). It has been noted that generous application of opioids can result in improved MEP monitoring owing to the reduction of spontaneous muscle contraction and lowering of the MAC for other anesthetic agents. With regards to the EEG, opioids produce a mild slowing into the delta range without effect on amplitude. Opioids will not produce burst suppression or an isoelectric EEG even at the highest doses. Of particular importance, the development of remifentanil has revolutionized opioid use in TIVA. Remifentanil is an ultra-short-acting opioid with a half-life on the order of 5 minutes regardless of dose. This allows for very rapid titration of analgesia with little or no effect on emergence times, thus permitting high-dose opioid TIVA to minimize the dose of an associated sedative-hypnotic.

Ketamine Ketamine is one of the older anesthetic agents and has undergone a recent resurgence of use owing to the finding that it helps to alleviate postoperative pain and chronic pain states. Ketamine influences a variety of receptors and has the unique characteristic among anesthetic agents of enhancing EP responses, especially in the cortex and spinal cord (15). Whether given as single bolus at induction or as continuous infusion, ketamine can increase EP amplitude in SEP, MEP, and BAEP recording, making it an attractive agent for use dur-

ing NIOM. Additionally, the use of ketamine can produce larger amplitudes, with mild slowing into the theta range on the EEG, and there is anecdotal evidence that ketamine may be proconvulsant. The downside to ketamine use (and the reason ketamine fell out of favor prior to the last 5 years) is the occurrence of emergence delirium and dissociative hallucinations. Additionally, increase in intracranial pressure from enhanced cerebral blood flow due to ketamine makes it of limited use in neurosurgical patients with intracranial hypertension as well as in some other patient populations. Ketamine has been found particularly useful as a low-dose infusion (10 to 20 µg/kg/min) to supplement a propofol/opioid TIVA technique in procedures that require anesthetic-sensitive NIOM (e.g., MEP). The addition of low-dose ketamine to a propofol-based TIVA allows for a substantial reduction in propofol infusion doses and enhancement of EP responses while minimizing the undesirable side effects of ketamine. For procedures requiring NIOM techniques that are highly sensitive to the effects of anesthetics (e.g., transcranial MEP), the use of ketamine in the anesthetic armamentarium should be considered.

Etomidate Etomidate represents another contradiction to the general rule that anesthetic agents cause EP depression. Induction doses and continuous intravenous infusion enhance both MEP and SEP recordings (16). Etomidate has been used in the past as a component of TIVA during procedures that require anesthetic-sensitive NIOM (e.g., transcranial MEPs). Etomidate is also somewhat contradictory in its EEG effects; at low doses it may be somewhat proconvulsant, and it is occasionally used for ECT or epilepsy surgery; although at higher doses it may produce burst suppression. However, among its many unpleasant side effects, concerns have been raised regarding etomidate-induced adrenal suppression, which can occur with even single-bolus induc-

Husain 04

1/17/08

11:55 AM

Page 63

CH A P T ER 4: A ne s the tic Cons id e rations

tion doses (0.2 to 0.5 mg/kg). Increased mortality has been seen with prolonged infusion of etomidate, mainly in the intensive care setting (17). Nevertheless, etomidate remains valuable in cases where NIOM responses are difficult to obtain and otherwise may not be recordable.

Dexmedetomidine Dexmedetomidine is a relatively new agent used in human anesthesia. This selective alpha-2 agonist has seen widespread use in veterinary medicine and has found its way into intensive care units and operating rooms because of its desirable effects of sedation, analgesia, and sympatholysis without respiratory depression. Though increasing ancedotal reports are emerging, there are limited data on the effects of dexmedetomidine on NIOM; however, animal data suggest that there is little effect (18). It may be used as a low-dose infusion (0.2 to 0.5 µg/kg/hr) to augment any anesthetic technique, and it allows for the use of considerably less volatile or intravenous anesthesthesia or opioid. Its definitive role in anesthetic techniques for highly sensitive NIOM remains to be determined.

Paralytics Neuromuscular blockers exert their effect by blocking acetylcholine at the nicotinic receptor in the neuromuscular junction. They have no effect on monitoring modalities that are not derived from muscle activity (e.g., EEG, BAEPs, and SEPs). They will completely negate MEP monitoring if intense neuromuscular blockade is utilized. However, employing partial blockade will allow substantial reduction in patient movement with testing, improved surgical retraction, and favorable MEP monitoring. There are many ways to monitor the amount of neuromuscular blockade; the most common is the “train of four” (TOF) technique. It consists of measuring muscle responses, or compound muscle action potentials, after four 2-Hz peripheral

•

63

nerve stimuli. MEP monitoring is acceptable when neuromuscular blockade is maintained at a TOF of two responses. In using MEP monitoring, it is important for the neurophysiologist and surgeon to know whether the patient is paralyzed. If the patient is not paralyzed, MEP stimulation must be done at times when patient movement is acceptable. If the patient is paralyzed, there are likely to be brief periods when MEP responses are not recordable owing to intense paralysis; it is then imperative to communicate when a neuromuscular blocking agent is redosed. However, either practice, paralysis or not, is acceptable; the main principle is, again, effective and open communication with all parties in the surgical suite.

ANESTHETIC TECHNIQUES A variety of anesthetic techniques are acceptable for use during NIOM; the type of anesthetic should be tailored to the type of NIOM and the requirements of the surgical procedure. There are, however, a few general principles. First, the least amount of anesthetic agent necessary should be utilized as long as there is little possibility of awareness or discomfort on the part of the patient. The liberal use of opioids can allow for a significant decrement in MAC. Second, the more stable an anesthetic dose can remain for the duration of the case, the less likely that the anesthetic agent might be contributing to intraoperative changes in NIOM waveforms. Supplementation of baseline anesthetic drugs with opioids or less NIOM-offending agents can be made at times of more intense surgical stimulation. Overall, there are essentially four classes of NIOM based on how easily the monitoring technique is ablated by anesthetic agents. As the relative sensitivity of NIOM to anesthesia increases, the anesthetic technique should be adjusted to maximize the least offending agents. Each group and its anesthetic implications are discussed below.

Husain 04

64

1/17/08

•

11:55 AM

Page 64

SE C T I ON I : B a s i c P r i n c iple s

Relative Insensitivity NIOM that is relatively insensitive to anesthetic agents in general includes BAEPs and SEPs recorded from the epidural space. With these monitoring methods, nearly all anesthetic practices can be used with the understanding that the general objective is to maintain a constant level of anesthesia supplemented with intermittent opioid dosing to control increased surgical stimulus. Of course the least amount of anesthetic necessary to ensure amnesia and analgesia should be used. Generally all patients have baseline EPs, so that once in the operating room, deviation from that baseline can be assessed. If needed, anesthetic level or technique can then be adjusted to refine NIOM recordings.

Sensitivity to Anesthetics without Sensitivity to Paralysis NIOM that is not negated by neuromuscular blockade but is sensitive to anesthetic agents includes SEP monitoring. Care must be taken to minimize offending anesthetic agents and optimize non-anesthetic variables (i.e., temperature). Generally, volatile or intravenous anesthesia is acceptable if relatively low doses are maintained (0.5 MAC for anesthetic gases or less than 80 µg/kg/min of propofol). The use of neuromuscular blockade in this situation allows for a modest decrement in anesthetic dose, as patient movement and relaxation then become improbable. However, care must be taken that anesthetic dose is not so low as to permit patient recall or discomfort.

Sensitivity to Paralytics

Relative Sensitivity

Forms of NIOM that are sensitive to neuromuscular blockade include all monitoring that requires elicitation of muscle action potentials (i.e., electromyography, MEP, spinal reflex testing, etc.). For these cases, if very fine control of the amount of neuromuscular blockade can be maintained through vigilant monitoring and drug dosing, neuromuscular blocking agents can be employed. Otherwise they should be entirely avoided once the patient has been intubated. In fact, there are some practices that utilize intraoperative neuromuscular blockade reversal when critical monitoring periods approach. In general, with the exception of MEP recording, which is exquisitely sensitive to anesthetic technique, other forms of anesthetic agents are acceptable. For cases that rely on an unparalyzed patient, relatively “deep” anesthesia (e.g., high doses of anesthetic agents) can be used to offset lack of patient paralysis, allowing optimal surgical conditions of immobility and relaxation while maintaining the integrity of NIOM. However, the general principle of stable, though relatively high, anesthetic dose should be maintained.

The need for MEP monitoring can initiate some of the more challenging anesthetic issues. Designing an anesthetic technique to optimize MEP monitoring adds to an already complex surgical procedure. A TIVA technique that limits the amount of sedative-hypnotic agent (i.e., propofol, barbiturate) is usually required. Limiting the dose of sedative-hypnotic to allow for optimal response recording of NIOM requires the use of a second agent, usually opioid, to supplement and augment the anesthetic properties. For instance, using a propofol-based anesthetic requires the addition of opioid, ketamine, or dexmedetomidine infusion to allow a much smaller dose of propofol to be administered. Additionally, if neuromuscular blockade is used, it must be tightly controlled so that profound paralysis does not preclude MEP responses from the muscles. It is not uncommon for the patient to be unparalyzed during critical monitoring portions of the procedure. Therefore the anesthesiologist is often faced with an unparalyzed patient, whose monitoring requires relatively low doses of an anesthetic, and whose airway and accessibility is often remote. One current

Husain 04

1/17/08

11:55 AM

Page 65

CH A P T ER 4: A ne s the tic Cons id e rations

practice is to utilize high-dose remifentanil infusion to supplement a low-dose propofolketamine based anesthetic. This allows very low dose propofol (20 to 30 µg/kg/min), which has minimal effects on MEP responses, to be offset by low-dose ketamine (10 to 20 µg/kg/min), which enhances MEP responses, and an amount of remifentanil that keeps the patient motionless and relaxed.

2.

3.

4.

CONCLUSIONS In developing an anesthetic plan, the type of NIOM is often as important a consideration as the type of surgical procedure. The crucial factor for a successful procedure is open and candid communication between the operating room staff, neurophysiologist, anesthesiologist, and surgeon. The majority of problems with intraoperative monitoring arise when operating room communication does not allow for each individual to have a clear understanding of the actions of each of the other members. When everyone involved in the procedure is knowledgeable about reasonable expectations and aware of the current situation, the patient benefits from an operating team that is poised and fluid in its execution. With that understanding, it is imperative for the neurophysiologist to understand the limitations produced by an anesthetic and for the anesthesiologist to understand the effects of certain medications on monitoring. Without that fundamental knowledge, there can be little coordinated activity between the two parties, resulting in ineffective monitoring. However, with the knowledge of the basic effect of a given anesthetic agent on monitoring modalities, nearly any anesthetic technique can be administered safely and effectively with all types of monitoring.

5.

6.

7.

8.

9.

10.

11.

12.

REFERENCES 1. James ML, Husain AM. Brainstem auditory evoked potential monitoring: when is change

13.

•

65

in wave V significant? Neurology 2005;65: 1551–1555. Sloan TB. Evoked potentials In: Albin MS, ed. A Textbook of Neuroanesthesia with Neurosurgical and Neuroscience Perspectives. New York: McGraw-Hill, 1997:221–276. Stekker MM, Escherich A, Patterson T, et al. Effects of acute hypoxemia/ischemia on EEG and evoked responses at normothermia and hypothermia in humans. Med Sci Monit 2002;8:CR223–CR228. Dolan EJ, Transfeldt EE, Tator CH, et al. The effect of spinal distraction on regional spinal cord blood flow in cats. J Neurosurg 1980;53: 756–764. Grundy BL, Heros RC, Tung AS, et al. Intraoperative hypoxia detected by evoked potential monitoring. Anesth Analg 1981;60: 437–439. Nagoa S, Roccaforte P, Moody RA. The effects of isovolemic hemodilution and reinfusion of packed erythrocytes on somatosensory and visual evoked potentials. J Surg Res 1978; 25:530–537. Mackey JR, Hall JW III. Sensory evoked responses in head injury. Central Nerv Syst Trauma 1985;2:187–206. McGirt MJ, Woodworth GF, Brooke BS, et al. Hyperglycemia independently increases the risk of perioperative stroke, myocardial infarction, and death after carotid endarterectomy. Neurosurgery 2006;58:1066–1073. Gugino LD, Aglio LS, Segal NE, et al. Use of transcranial magnetic stimulation for monitoring spinal cord motor paths. Semin Spine Surg 1997;9:315–336. Deletis V. Intraoperative monitoring of the functional integrity of the motor pathways. Adv Neurol 1993;63:201–214. Drummond JC, Todd MM, U HS. The effects of high dose sodium thiopental on brainstem auditory and median nerve somatosensory evoked responses in humans. Anesthesiology 1985;63:249–254. Sloan TB, Fugina ML, Toleikis JR. Effects of midazolam on median nerve somatosensory evoked potentials. Br J Anaesth 1990;64: 590–593. Kalkman CJ, Drummond JC, Ribberrink AA. Effects of propofol, etomidate, midazolam,

Husain 04

66

1/17/08

•

11:55 AM

Page 66

SE C T I ON I : B a s i c P r i n c iple s

and fentanyl on motor evoked responsesto transcranial electrical or magnetic stimulation in humans. Anesthesiology 1992;76:502–509. 14. Angel BA, LeBeau F. A comparison of the effects of propofol with other anesthetic agents on the centripetal transmission of sensory information. Gen Pharmacol 1992;23: 945–063. 15. Schubert A, Licina MG, Lineberry PJ. The effect of ketamine on human somatosensory evoked potentials after high frequency repetitive electrical stimulation. Electroencephalogr Clin Neurophysiol 1998;108:175–181.

16. Kochs E, Treede RD, Schulte AM, Esch J. Increase in somatosensory evoked potentials during induction of anaesthesia with etomidate. Anaesthetist 1986;35:359–364. 17. Ledingham IM, Watt I. Influence of sedation on mortality in critically ill multiple trauma patients. Lancet 1983;1:1270. 18. Li BH, Lohmann JS, Schuler HG, Cronin AJ. Preservation of the cortical somatosensoryevoked potential during dexmedetomidine infusion in rates. Anesth Analg 2003;96: 1150–1160.

Husain 05

1/17/08

11:57 AM

Page 67

5

Billing, Ethical, and Legal Issues*

Marc R. Nuwer

P

system, these codes allow for reimbursement proportional to the time, effort, equipment, supplies, and risk involved in a procedure. Each specific procedure has its own code number, assigned in the American Medical Association’s Current Procedural Coding (CPT). CPT specifies more than 7,500 medical, surgical, and diagnostic procedures. Cost accounting for each CPT code is determined by the typical work, equipment, supplies, etc., involved in carrying out that procedure. These costs are systematically summarized in the Resource Based Relative Value System (RBRVS), which is used to set the Relative Value Units (RVU) for each CPT. Most U.S. healthcare insurance carriers use RVUs for payment. Other organizations use physician work RVUs to judge the annual work productivity of group physicians. CPT codes used for most NIOM are presented in Table 5.1. The vast majority of cases use the hourly NIOM code CPT 95920. Other specialty codes are used for certain electroencephalograms (EEG) and functional mapping procedures. The codes also allow for a systematic set of additional policies. The most relevant ones are those that specify the degree of supervision required for each type of procedure.

rofessional administrative policies and procedures assure patient safety, efficient organization, and clear communication. Well-defined policies discourage self-interested motives and clarify limits. Public policies on health care include laws and regulations about billing, coding, reimbursement, staffing, devices, and liability, among others topics. This chapter describes some public policy issues that apply to neurophysiologic intraoperative monitoring (NIOM). Issues pertaining to billing apply specifically to the United States. Other ethical and legal issues are general principles and applicable to NIOM wherever monitoring is conducted. Of course, in some legal jurisdictions, other locally specific laws and regulations may apply too.

CODING AND BILLING These coding and billing rules apply to the United States. This coding system specifies the particular work conducted on a particular patient. In the U.S. medical reimbursement * This chapter is based in part on Nuwer MR. Regulatory and medical-legal aspects of intraoperative monitoring. J Clin Neurophysiol 2002;19:387–395.

67

Husain 05

1/17/08

68

•

11:57 AM

Page 68

SE C T I ON I : B a s i c P r i n c iple s

TABLE 5.1 Intraoperative Neurophysiology CPT Codes 95920

Intraoperative neurophysiology testing, per hour

95829

Electrocorticogram at surgery

95955

Electroencephalogram during nonintracranial surgery (e.g., carotid surgery)

95961

Functional cortical and subcortical mapping by stimulation and/or recording of electrodes on brain surface, or of depth electrodes, to provoke seizures or identify vital brain structures Initial hour of physician attendance

95962

Each additional hour of physician attendance

Code 95920 (Intraoperative Neurophysiology Testing) This code is billed once for each hour of routine NIOM. Time begins when the physician is called to the operating room and ends when he or she leaves. However, the time taken to interpret the baseline testing for each primary base procedure is excluded, since those are billed separately. Time is rounded to the nearest hour. For remote online monitoring, the code is used for each hour monitored. Supervision rules require direct supervision, which may be done remotely online. Direct supervision ordinarily requires the clinical neurophysiologist to be present nearby in the building, readily available to come into the surgery suite on short notice when a problem or need arises. Remote on-line monitoring allows for the equivalent done at a distance, with available interaction by phone or instant messaging. In either case, the clinical neurophysiologist still must be available to respond on short notice if problems occur. Code 95920 is used along with base codes that specify what modalities were monitored or tested. The base (primary) procedures codes

are billed once per procedure. The allowable base procedure codes for code 95920 are listed in Table 5.2. Only these base codes are allowed. At least one must be used for any procedure when code 95920 is used. More than one code may be used when more than one modality is monitored. For example, monitoring resection of an acoustic neuroma may include upper extremity somatosensory evoked potentials, facial nerve electromyography (EMG), and brainstem auditory evoked potentials. Those base codes are 95925, 95868, and 92585, and they are coded with one unit of service each. Code 95920 is billed one unit of service for each monitoring hour TABLE 5.2 Primary Procedures Allowed as Base Codes for Code 95920 92585

Auditory evoked potential

95822

EEG recording in coma or sleep

95860

Needle EMG One extremity

95861

Two extremities

95867

Cranial nerve muscles, unilateral

95868

Cranial nerve muscles, bilateral

95970

Limited

95900

Nerve conduction study, each nerve Motor

95904 95925

Sensory Somatosensory evoked potential In upper limbs

95926 95927 95928

In lower limbs In trunk or head Central motor evoked potential In upper limbs

95929

In lower limbs

95930

Visual evoked potential

95933

Blink reflex

95934

H reflex Gastrocnemius/soleus muscle

95936 95937

Muscle other than gastrocnemius/ soleus muscle Neuromuscular junction testing (repetitive nerve stimulation)

Husain 05

1/17/08

11:57 AM

Page 69

CH A P T ER 5: Billing, Ethical, and Le gal Is s ue s

after excluding the time taken to interpret the baseline portions of each monitoring modality. Other procedures can be performed on the same day. For example, laryngeal EMG is coded as CPT 95865. Although it is not a base code for 95920, it can also be coded in addition to other codes if laryngeal EMG was monitored.

Code 95829 (Electrocorticography) Code 95829 is used for electrocorticography (ECoG). The intraoperative evaluation and medical decisions during an ECoG go well beyond simple monitoring. Code 95920 is used generally to screen for any changes compared to baseline where changes lead to raising an alarm. By contrast, in ECoG, the physician identifies regions to be removed and to be left intact during cortical resections. The difference between monitoring (95920) and deciding what to remove (ECoG) justifies a separate procedure code. Because this procedure requires medical decision making of a high degree, this code and service require the personal supervision of the physician in the operating room.

Code 95955 (Nonintracranial EEG) For historical reasons, intraoperative EEG during nonintracranial surgery has its own code, 95955. This can be used for EEG during carotid endarterectomy or cardiothoracic surgery. Traditionally this was a “general supervision” service. In that traditional supervision system, an advanced or specialized technologist conducted the EEG monitoring without immediate physician supervision. When a physician provides a greater degree of supervision throughout the procedure, code 95920 can be used, together with the base procedure code 95822 for intraoperative EEG monitoring. The difference between 95955 and 95920 is the degree of supervision.

•

69

Codes 95961 and 95962 (Cortical and Subcortical Localization) Codes 95961 and 95962 are used when localizing cortical or hemispheric function. This can be localizing motor or language cortex or when placing deep brain stimulator (DBS) electrodes. Code 95961 is used for the first hour. Code 95962 is used for each additional hour. These same codes are used for testing cortical functions in monitoring epilepsy patients with subdural grids. In the operating room, these codes are used in identifying language regions with direct cortical stimulation in an awake patient during a craniotomy.

Codes 95970 to 95979 (Neurostimulator Programming and Analysis) Neurostimulator programming and analysis use the same codes, 95970 to 95979, inside and outside the operating room. The several neurostimulator codes are listed in Table 5.3. The codes come in simple and complex types. A simple neurostimulator pulse generator/transmitter has three or fewer adjustable parameters. A complex neurostimulator pulse generator/transmitter is one capable of affecting more than three of the above. Modern DBS systems involve the complex codes. Code 95970 is used when one is just checking the integrity of an implanted neurostimulator without programming it. For DBS brain programming, code 95978 should be used for the first hour and 95979 for each additional 30 minutes. For vagus nerve stimulator programming, code 95974 should be used for the first hour and 95975 for each additional 30 minutes. In some surgical DBS placements, codes 95961 to 95962 are used to identify the correct locations for the electrodes. Then, the 95870 code is used to assure that the DBS stimulating device is working properly.

Husain 05

1/17/08

70

•

11:57 AM

Page 70

SE C T I ON I : B a s i c P r i n c iple s

TABLE 5.3 Neurostimulator CPT Codes 95970

Implanted neurostimulator analysis

95971

Simple neurostimulator programming

95974

Complex cranial nerve neurostimulator programming First hour

95975 95978

Each additional 30 minutes Complex brain stimulator programming First hour

95979

Each additional 30 minutes

STAFFING Most states and hospitals require a licensed physician to supervise each technologist. Such public policies limit the practice of nonphysician healthcare providers.

Interpreter Staffing A technologist cannot provide NIOM without any supervision. Many lack the suitable skills, knowledge, ability, training, and experience to provide interpretations and respond to the more challenging moments of monitoring. Many technologists have insufficient medical knowledge to advise the surgeon on the medical meaning of signal changes. Based on such limitations, most hospitals and states require physician supervision. Surgeons generally cannot supervise the technologist for NIOM. Most surgeons lack suitable skills, knowledge, abilities, training, and experience in clinical neurophysiology to assure high-quality monitoring services,which can lead to mistaken interpretations. They lack an understanding of the technical procedures involved, normal variations and peak identification procedures, artifacts and other technical problems, and the meaning of changes when they occur. The National Correct Coding Initiative (NCCI) precludes billing for both the surgery and the NIOM by the same physician.

A monitoring physician must be available to intervene as needed when problems arise in any case. The monitoring physician also must be able to pay close attention to the individual cases when monitoring simultaneous procedures. For those reasons a limit of three simultaneous cases per monitoring physician is recommended (1). A nonphysician healthcare provider sometimes does supervise intraoperative monitoring technologists. Any provider must have advanced skills, knowledge, ability, and training, as well as extensive experience to assure quality services and patients’ safety. Measuring that can be difficult. The judgment must be made by reasonably independent observers and based upon objective guidelines and standards. There is in place a professional process for assessing physicians’ credentials for hospital privileging. These processes can serve as a model or outline of processes to assess nonphysician providers. Each hospital’s medical staff office must assure that each professional meets appropriate standards to practice in that facility. Privileging determines whether an individual practitioner at a given institution should be allowed to practice and for what procedures. The institution checks for medical board and DEA certification, professional disciplinary actions, licensure, training, experience, proof of insurance, malpractice claims, and criminal convictions. Individual proctoring checks that actual practice performance currently meets local community standards. Credentialing evaluates an individual’s specific fields of practice and particular techniques for which he or she should be allowed to practice at the institution in question. This examination must be conducted even more thoroughly for any nonphysician than for a physician because there is no state licensure agency also overseeing such practitioners.

Technologist Staffing Not all technologists are suitable for NIOM. Registration by the American Board of Electroencephalographic and Evoked Potentials

Husain 05

1/17/08

11:57 AM

Page 71

CH A P T ER 5: Billing, Ethical, and Le gal Is s ue s

Technologists (ABRET) as an R. EEG T. (registered EEG technologist), R. EP T. (registered evoked potential technologist), and Certificate in Neurophysiologic Intraoperative Monitoring (CNIM) provides one type of check on an individual’s NIOM knowledge. More is needed. Hospitals’ neurodiagnostics departments should evaluate each technologist for the his or her skills, knowledge, abilities, training, and experience relevant to NIOM. The lab medical director should establish suitable policies and procedures. Each technologist should be privileged to perform just the NIOM procedures for which he or she is suitably trained. A technologist well skilled in one type of procedure is not necessarily qualified to conduct all types of procedures. Privileging should be subject to annual review and renewal. For example, NIOM without immediate supervision might be restricted to technologists with 3 to 5 years of experience with that technique outside the operating room. An extended period of proctoring would involve a period of immediate supervision followed by a period with less supervision until the individual becomes fully qualified. Some procedures require a physician to be in the room. These include ECoG, functional cortical localization, and other testing for deliberate decisions about what to resect or spare. A technologist can conduct more routine monitoring without supervisory backup in the room. A physician is commonly available on-line (2). Typically, monitoring technologists are given specific policies and procedures outlining their responsibilities for specific kinds of monitoring services. These policies and procedures can script the specific instances in which the technologist is expected to report adverse changes to the surgical team or to physician or supervisory backup.

MEDICAL-LEGAL ISSUES Lawsuits can occur after any surgical procedure, even if the monitoring community’s practice standards were met. Monitoring does

•

71

not necessarily eliminate adverse neurologic events; it just reduces their incidence (3). Yet errors do occur in medical or surgical care. The Institute of Medicine reported that between 48,000 and 98,000 Americans die each year after incidents associated with medical errors (4). Such errors include prescribing or dispensing the wrong medication. Many other problems occur too. Intraoperative monitoring is a situation where one can encounter communications errors. The NIOM team needs to keep the surgeon aware of the state of nervous system monitoring. In turn, the surgeon needs to know the role of monitoring and realize the possibility of neurologic injury despite monitoring. Monitoring should be conducted by welltrained staff. Recording quality must be maintained. When quality is poor or results become irreproducible, the surgeon should be told that the monitoring cannot be trusted. Documentation includes labeling the pages with the patient’s name and the time the tracings were collected. A flow sheet should be maintained during the case, summarizing the status of the peaks and any comments to the surgeons.

CONCLUSIONS Monitoring teams should adhere to proper professional quality and standards. Services should function with appropriate policies on supervision, staffing, privileging, credentialing, and certifying clinical neurophysiologists and monitoring technologists. These should assure each individual’s skills, knowledge, abilities, and training relevant to monitoring. Good record documentation and clear communications are also important. The goals remain to serve the patients first, do no harm, and protect the trust that the public places in us as professional caregivers. NIOM can protect patients, enhance good outcomes, and encourage more thorough surgical procedures.

Husain 05

72

1/17/08

•

11:57 AM

Page 72

SE C T I ON I : B a s i c P r i n c iple s

REFERENCES 1. Nuwer MR. Intraoperative monitoring of the spinal cord. Clin Neurophysiol, 2008;119(2): 247–248. 2. Nuwer JM, Nuwer MR. Neurophysiologic surgical monitoring staffing patterns in the USA. Electroencephalogr Clin Neurophysiol 1997;103(6):616–620. 3. Nuwer MR, Dawson EG, Carlson LG, et al. Somatosensory evoked potential spinal cord

monitoring reduces neurologic deficits after scoliosis surgery: results of a large multicenter survey. Electroencephalogr Clin Neurophysiol 1995;96(1):6–11. 4. Kohn LT, Corrigan JM, Donaldson MS, eds. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press, 1999.

Husain 06

1/17/08

11:58 AM

6

Page 73

A Buyer’s Guide to Monitoring Equipment

Greg Niznik

that all testing was done with the subject seated in a Faraday cage, basically a room completely lined with copper wire mesh that was grounded to both the electrical circuit ground and the signal averager. The Faraday cage/room was borrowed from the neurophysiology laboratory, where intra-and extracellular recordings of neural preparations were made. It was only big enough to hold a recliner and a rack for the amplifiers. The amplifiers were so unstable that they had to be reset and have the DC offset zeroed before each trial. The stimulators, signal averager, oscilloscope, and plotter were rack-mounted just outside the Faraday cage. In the early 1980s, the first multichannel evoked potential (EP) machines became available. These were portable (at least they had wheels) and had integrated, fairly stable amplifier and stimulator systems. More importantly, they were true digital machines that could display and store banks of data. These machines were, for the most part, large and cumbersome towers or consoles weighing several hundreds of pounds. They were a far cry from today’s laptop-based systems but they represented a giant leap forward in technology. Having a four- or eight-channel machine with at least two electrical, an auditory, and visual stimulators made the development of NIOM techniques possible. Having

T

he title of this chapter is a bit of a misnomer, as any modern commercially available NIOM machine will, in fact, fulfill its stated purpose, which is to monitor the nervous system during surgery. Therefore this chapter provides a review of the minimum requirements of NIOM equipment as outlined in the 2006 ACNS guidelines (1). These guidelines and their importance as they relate to NIOM machines are discussed. By having a better understanding of the workings of the NIOM machine, the reader will be better able to make an informed decision in selecting a NIOM machine for his or her practice and be better able to utilize it to its full capability.

A SHORT HISTORY OF A BURGEONING TECHNOLOGY In the early days of NIOM, most equipment was designed and built by the investigator, either from the ground up or from individual stand-alone components. For example, the author’s first clinical evoked potential (EP) system consisted of a one-channel signal averager, battery-powered amplifiers, a constant-voltage stimulator, an analog oscilloscope to display the data, and a pen plotter to provide hard copy of the data. The system was so susceptible to extraneous noise 73

Husain 06

74

1/17/08

•

11:58 AM

Page 74

SE C T I ON I : B a s i c P r i n c iple s

stable amplifiers and good filters made the recording of microvolt signals in the electrically hostile environment of the operating room feasible. These machines also had paradigms for signal rejection, which ensured that electrocautery or movement artifact did not destroy the averaged response acquired over the preceding several minutes. Many had available programming languages that facilitated some very basic automation. These early machines performed only EP testing. If cortical brain function had to be monitored along with EPs, a separate electroencephalography (EEG) machine had to be brought into the operating room. Therefore the space requirement in the operating room for NIOM was not insignificant. As personal computers became more powerful and as the technology for shrinking digital components improved, more powerful and compact NIOM equipment came on the market. With these, NIOM techniques improved in leaps and bounds. Towers became carts and the number of channels increased, so that eight channels was now the norm. Four electrical stimulators meant that the upper and lower spinal cord could be monitored simultaneously. Amplifiers became even more reliable and stable, reducing the downtime that was previously suffered by saturated amplifiers. True interleaving became available, where data could be collected from several stimulators simultaneously, greatly improving the safety and efficacy offered by NIOM. Prior to this, stimulators were chained, which meant that a set of averages for one stimulator was completed before the next stimulator was activated. Now, with interleaving, each stimulator collected its data simultaneously, so that a complete set of averages was performed for all modalities at once. NIOM machines now were able to perform not only EP testing but also EEG, electromyography (EMG) and nerve conduction studies (NCS) all in one machine. This greatly reduced the amount of space that the NIOM team took up in the operating room. Digital

filtering and smoothing made it possible to improve the fidelity and readability of NIOM waveforms. Today truly portable laptop-based machines are available; 16- and even 32-channel machines are becoming the norm. Automation of data collection, ability to monitor numerous modalities simultaneously, and extremely stable amplifiers coupled with noise-reduction paradigms have taken NIOM technology and techniques to a whole new level. These advances reduce the time that the technologist spends troubleshooting, allowing him or her to concentrate on monitoring the function of the nervous system. Creative ways of displaying and trending data makes identifying significant changes easier and allow earlier intervention. The advent of on-line, real-time supervision of the technologist’s work by the neurologist or neurophysiologist has greatly increased the efficacy of NIOM, making surgery safer for the patient. NIOM is still in its infancy. The technology and techniques used have certainly grown by leaps and bounds over the last 30 years. As technology continues to improve, it is exciting to speculate where the field will be over the next several decades. One thing is certain: improvements in technology and technique will make surgery on and around the nervous system more efficacious and safer.

BASICS OF THE NIOM MACHINE An NIOM machine consists of three basic parts: some type of stimulator to activate a portion of the nervous system, an amplification system, and a central computer system to digitize, analyze, display, and store the resulting waveform. Each of these parts is discussed separately.

NIOM Stimulators NIOM stimulators are used to activate that portion of the nervous system which is at

1/17/08

11:58 AM

Page 75

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt

Electrical Stimulators The NIOM machine has electrical stimulators to activate neural structures during somatosensory EPs (SEP), transcranial electrical motor EPs (tceMEP), evoked EMG, and NCS. Typically square-wave pulses of varying intensity, duration, and frequency are delivered. The NIOM machine’s central computer controls these functions. In general an NIOM machine will have multiple stimulators available. A minimum of four stimulators is needed to provide simultaneous upper and lower extremity SEPs. Additional electrical stimulators are often available to handle special functions, such as low-power stimulators for direct nerve activation in situ and highvoltage stimulators for tceMEPs.

•

75

Stimulus Duration The NIOM electrical stimulator must be able to vary its stimulus duration to meet the needs of each specific surgery. Stimulus duration should be variable from 0.5 to 500 µs. Energy delivered to a nerve is the product of the stimulus intensity and pulse width. This product is known as power (P), which is represented as the area under the stimulus curve (Figure 6.1A). The same power can be delivered to a nerve by using a high intensity and short pulse width or by using a lower intensity and longer pulse width (Figure 6.1B). The maximum power that the NIOM machine can deliver must be limited so that burns do not occur.

A

Intensity

risk during surgery. In clinical EP, these are electrical stimulators that activate the somatosensory system, auditory stimulators that activate the auditory system, and visual stimulators that activate the visual system. Monitoring visual EPs is unreliable in the operating room. They are not used routinely; therefore intraoperative visual EPs are not dealt with in this chapter.

Power = Intensity x Pulse Width

Pulse Width

Stimulus Intensity

B

=

2X Intensity

An NIOM machine must be versatile in its ability to deliver an electrical stimulus to activate nerves. It must be powerful enough to activate deep neural structures through transcutaneous stimulation, as when the nerves in the arms and legs are being stimulated for SEPs. It must also be able to deliver the fine stimuli needed when neural tissue is being teased away from a tumor mass. It must be able to deliver the stimulus with fidelity and have fine control over the intensity delivered. This control should be in at least 10% increments of the stimulus delivered. For example, for low-current stimuli of less than 1 mA, control over the stimulus intensity in 0.1mA increments is necessary. For stimuli in the tens of mAs, 1 mA increments suffice.

Intensity

Husain 06

Pulse Width 1/2 x Pulse Width

FIGURE 6.1 A. The power of an electrical stimulus is the product of the intensity and pulse width. Power is represented by the area under the square-wave pulse. B. The same stimulus power can be delivered by varying the pulse width and stimulus intensity. These two stimuli have the same power; the first has a long pulse width and low intensity, the second has a short pulse width and high intensity.

Husain 06

76

1/17/08

•

11:58 AM

Page 76

SE C T I ON I : B a s i c P r i n c iple s

Stimulus duration has a direct effect on the presence and magnitude of the stimulus artifact. Using a short stimulus duration, especially when stimulating and recording over short segments, can help minimize the presence of stimulus artifact. As shown in Figure 6.2, the electrical properties of skin can be modeled using electrical components. Subcutaneous tissue acts as a capacitor, which charges and discharges with every electrical pulse. The greater the pulse width of the delivered electrical pulse, the longer the charge and discharge of the capacitor. This then leads to a larger capacitive stimulus artifact in the waveform. By keeping the pulse width of the stimulator relatively short, the capacitive effect of the skin is minimized, thus minimizing the capacitive artifact that may impinge on the recorded signal. Some equipment manufacturers provide stimulators that, instead of delivering a square wave, deliver a biphasic pulse in which the amplitudes of the positive and negative portions of the stimulus can be var-

Ra

Rt

Ra

Ra

Ct

Ra

FIGURE 6.2 A simplified equivalence circuit of current flow across a membrane can be used to model how a stimulus current flows across the skin and subcutaneous tissue. The circuit consists of resistance to axial flow of current (Ra); that is, current flow along the surface of the skin and subcutaneous tissue. Current flowing through skin and subcutaneous tissue is represented by a resistor and capacitor in parallel (Rt, Ct).

ied by the technologist. The biphasic pulse minimizes the buildup of charge across the skin, therefore minimizing the presence of a capacitive stimulus artifact.

Constant Voltage vs. Constant Current The NIOM machine should be able to deliver both constant-current and constantvoltage stimuli. Generation of an action potential in a neuron is initiated by depolarization of the relative charge across it membrane. By passing a negative current across the outer surface of the neuronal membrane, the inside of the neuron becomes relatively more electrically positive until threshold is reached and the all-or-none cascade of events known as the action potential leads to propagation of an electrical signal along the neuron. In order to ensure that activation of a nerve is consistent for each stimulus delivered during repetitive stimulation, the current delivered at each pulse should be held constant. Therefore, in general, constant-current stimulation is the preferred method of electrical stimulation during NIOM. As explained in Ohm’s law, V = IR, voltage and current are related by the resistance within the stimulation circuit. For purposes of NIOM, this resistance is primarily skin resistance. Constant-current stimulation varies the stimulus voltage in order to ensure that the required and requested current is delivered at each pulse regardless of changes in skin resistance. Two exceptions are detailed in the literature where constant voltage is the preferred method of stimulation during NIOM procedures. These are during stimulation of cranial nerves during tumor resection and in tceMEP. Some authors have suggested that constantvoltage stimulation is preferred over constantcurrent stimulation when nerves in a wet field are stimulated, as it overcomes the problem of current shunting through fluid (2). When a nerve immersed in fluid is being stimulated, current flows through the path of least resistance, primarily through the fluid, and there is

Husain 06

1/17/08

11:58 AM

Page 77

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt

no way of determining how much current actually reaches the nerve. Furthermore, current shunting varies depending on how wet or dry the field is and how much is shunted through the adjacent fluid. It has been suggested that using constant-voltage stimuli should ensure adequate current delivery across a nerve regardless of whether the field is dry or wet. This theoretically would ensure that adequate depolarization would occur during all phases of surgery and under all stimulating conditions.

Transcranial Electrical Motor EP Until a few years ago, tceMEP stimulators were experimental and not approved by the FDA for use in routine clinical settings. Use of stand-alone stimulators or using a machine’s integrated constant-voltage stimulator to stimulate the motor cortex transcranially was considered an off-label use of this equipment. In late 2002, Digitimer Ltd. received the first FDA approval for their D185 stimulator for use as a transcranial electric stimulator. Since then other manufacturers have acquired FDA approval for their transcranial electrical stimulators. These subsequent approvals are based on the stimulus output specifications of the Digitimer D185. The D185 is a constant-voltage stimulator with a maximum output voltage of 1,000 V. Its maximum current output is 1.5 A. It has a fixed pulse width of 50 µs and a fast rise time of 0.1 A/ µs. TceMEP methods employ trains of stimuli to activate the motor cortex. The D185 delivers up to nine pulses per train at an interstimulus interval of 1.0 to 9.9 ms in 0.1 ms increments. Prior to using either a stand-alone tceMEP stimulator or a tceMEP stimulator integrated into a NIOM machine, one must determine whether the stimulator is FDAapproved for transcranial stimulation. If it is not, its use as a tceMEP stimulator is considered off-label. The use of a medical device for an off-label application may involve a series of regulatory procedures specific to individual institutions. At the very least, it may involve

•

77

making the surgeon aware that the tceMEP stimulator is not approved for transcranial stimulation. The surgeon deeming it necessary for the surgery prescribes its use and accepts responsibility for its safe and prudent operation. However, the institution may hold that off-label use of a stimulator requires Institutional Review Board (IRB) approval under a research protocol, that the patient be informed of the potential risks and benefits of this investigational procedure, and that the patient’s consent be obtained before this equipment is used. The bottom line is that one must check with the given institution as to what the requirements are for off-label use of medical equipment.

Auditory Stimulation Sound and the Decibel Scale Before embarking on a discussion of the specific properties needed in NIOM auditory stimulators, a clear explanation of the way sound is measured must be presented. The decibel, the often misunderstood unit of measurement of sound, pressure, power, and voltage is discussed first. This discussion of the decibel is also relevant to discussions of other aspects of the NIOM machine; for example, in describing how efficiently analog filters eliminate unwanted noise. The decibel (dB) is 1/10 bel, a basic unit of sound, named after Alexander Graham Bell. The decibel describes the relative strength of a sound. In other words, it describes how loud a sound is relative to another sound. Therefore if a sound has an intensity of 80 dB, its intensity is 80 dB above some reference value (i.e., the subject’s threshold of hearing, the threshold of hearing of a normal population, or background noise). It is always a comparison. What is perceived as sound is the result of a wave of air pressure that displaces the eardrum and is subsequently transduced into an electrical signal. Therefore sound intensity is in reality a measurement of air pressure on

Husain 06

78

1/17/08

•

11:58 AM

Page 78

SE C T I ON I : B a s i c P r i n c iple s

the eardrum. In terms of sound intensity, a decibel, then, is a measure of relative sound pressure. A 1-dB change in sound intensity approximates a just perceivable change in sound level by the human ear under ideal conditions. The ear perceives sound in a nonlinear fashion; in fact, it perceives sounds logarithmically. Therefore the decibel is a logarithmic measure. If the power of two sounds, such as those coming out of a loudspeaker, is measured, the decibel difference in their power is described by the following equation: dB 5 10 log10 (P2 / P1) where P 5 power. In this equation, doubling the power of a loudspeaker would increase it by approximately 3dB: dB 5 10 log10 (2) 5 10 (0.301) 5 3 However, in actuality, when sound level is measured, it is done in terms of pressure or voltage (through a microphone). Pressure or voltage is the square root of their power. Therefore, in describing a change in pressure or voltage, the formula for the decibel changes as follows: dB 5 20 log10 (p2 / p1) or dB 5 20 log10 (V1/V2) where p 5 sound pressure and V 5 voltage. In this equation, doubling the sound pressure would increase it by approximately 6dB: dB 5 20 log10 (2) 5 20 (0.301) 5 6

Sound Reference and the Decibel Scale As discussed above, the decibel scale is relative; the intensity of a particular sound is measured to a certain reference sound level. Three techniques of quantifying auditory stimuli are described in the ACNS guidelines for performing brainstem auditory EPs

(BAEPs) (1). These are decibels peak-equivalent sound pressure level (dB pe SPL), decibels above normal hearing level (dB HL), and decibels above sensation level (dB SL). One way in which these techniques differ is with respect to their reference value. dB pe SPL is a measure of the actual sound pressure delivered by the auditory stimulator. It is the unit used in calibrating auditory stimulators. The unit of pressure used in this measure is the micropascal, or dyne2/cm2. Sound pressure levels are compared to an arbitrary zero pressure level which is 20 micropascals, or 0.0002 dyne2/cm2. dB HL is a measure comparing sound intensity to the hearing threshold of normal, healthy young adults. Therefore a sound level of 6 dB HL is twice as loud as the threshold of a population of normal subjects hearing the same sound. This measure is usually default in NIOM machines and most commonly used during NIOM. dB SL is a measure using each patient as his or her own control. The reference point for this measure is the patient’s own hearing threshold. This is an important measure in clinical BAEPs in dealing with a patient with a hearing deficit. By using dB SL, the stimulus intensity can be matched to each ear, so that the same perceived intensity is delivered to both sides. NIOM machines should have the ability to provide stimuli in both dB HL and dB SL. dB pe SPL is impractical to use as a means of quantifying sound intensity in the operating room. It is imperative to know what measurement scales are available on one’s equipment and how to select the appropriate measure so that the expected stimulus intensity is delivered.

Auditory Stimulators The most common type of auditory stimulator used during NIOM is a piezoelectric transducer that delivers broadband clicks via a plastic tube and foam ear insert. These stimulators are commonly known as insert earphones. Traditional headphones are not

Husain 06

1/17/08

11:58 AM

Page 79

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt

recommended for intraoperative use as they are difficult to affix in such a manner that they do not become dislodged during surgery. It must be remembered that in using insert earphones, all BAEP waveform latencies will be delayed by about 1 ms. This occurs because of the time that it takes for the stimulus click to travel from the transducer through the tubing to the eardrum. Many NIOM machines have a feature that will automatically subtract this 1-ms delay if insert earphones are used. It behooves the operator to be aware of whether this feature is active when he or she is identifying BAEP waveform in the operating room. NIOM auditory stimulators typically deliver a broadband click produced by a 100µs electrical pulse to the transducer. The resultant click oscillates and can be up to 2 ms in duration. The polarity of the click should be selectable. Three types of polarity are available. The first, condensation, produces a sound wave that pushes the eardrum inward, away from the stimulator. Rarefaction produces a sound wave that pulls the eardrum outward, toward the stimulator. Alternating clicks, as the name implies, consist of alternating condensation and rarefaction stimuli. The intensity of the auditory stimuli, as discussed above, should be selectable. The stimulator should be capable of producing a broadband click with an intensity of up to 120 dB pe SPL. Most NIOM machines use dB HL as the default method of measuring intensity. Therefore the auditory intensity should be selectable from 0 to 100 dB HL in 1-dB steps. Often NIOM machines will allow the technologist to set a preoperative threshold level for the patient, in which case the stimulus intensity will be expressed as dB SL. The repetition rate for auditory clicks should be variable up to 200 Hz, although repetition rates in the range of 11 to 22 Hz are most common. Titration of the repetition rate to minimize the averaging time with acceptable degradation of signal fidelity is important, especially during surgery, where rapid assessment of the function of the auditory

•

79

pathway is imperative. Conversely, deceasing the repetition rate can often improve the fidelity of intraoperative BAEPs in dealing with patients with preoperative deficits. The availability of white noise masking is helpful in eliminating crossover of the auditory stimulus. Transmission of one-sided stimuli to the opposite ear by air or bone conduction can result in a BAEP that reflects activity in both ears. By delivering white noise at 60 dB SPL to the contralateral ear, this crossover component can be eliminated.

Amplification System The Operational Amplifier The operational amplifier (op-amp) is the basic building block of the amplification system of the NIOM machine. Its basic form is shown in Figure 6.3. It consists of two input poles and an output pole. It is a differential amplifier, meaning that its output depends on the nature of both of its inputs. One input, marked (–), is the inverting pole. It is also known as grid 1 (G1), the negative input, or the active input. The other pole, (+), is the noninverting pole, also known as grid 2 (G2), the positive input, or the reference input. The output signal consists of the difference between the input signals at the two poles (as seen in Figure 6.4). If the same signal is introduced at both the inverting and noninverting poles of the amplifier, no output signal will be seen. Signals introduced at the inverting pole

Inverting

(–)

Non-inverting

(+)

Output

FIGURE 6.3 The electrical symbol for the operational amplifier. It consists of two input poles, one inverting (– ve), the other noninverting (+ ve). The output signal is the arithmetic difference between the two input signals.

Husain 06

80

1/17/08

•

11:58 AM

Page 80

SE C T I ON I : B a s i c P r i n c iple s

A (–)

TABLE 6.1 Effect of Various Inputs on the Polarity of the Output Signal in an Op-Amp

Input

(+)

(–)

(+)

Output

B (–) (+)

C (–) (+) FIGURE 6.4 Examples of polarity convention in operational amplifiers. A. The same input signal is introduced at both inputs. The resulting output is a flat line. B. An upgoing signal is introduced at the inverting pole. The resulting output in downgoing. C. An upgoing signal is introduced at the noninverting pole. The resulting output is upgoing.

are, as suggested by the name, inverted at the output. If an upgoing signal is placed at the inverting pole and no signal at the noninverting pole, the output will be a downgoing signal. Signals introduced at the noninverting pole have the same polarity at the output. If an upgoing signal is introduced at the noninverting pole and no signal at the inverting pole, the resulting output will be an upgoing signal (Table 6.1). In NIOM applications where near-field potentials are recorded, the electrode over the generator site is placed in the inverting pole of the op-amp. The input to the noninverting pole is placed at a site away from the generator, where no activity will be recorded. Therefore electrically negative signals introduced into the op-amp will be displayed as an upgoing signal at the output. This is why the normal convention for most NIOM modalities is “negative up.” One exception to this during NIOM is the BAEP.

By convention, the BAEP is displayed “positive up.” Some NIOM machines automatically invert the BAEP display so that a negative signal is downgoing and a positive signal is upgoing (positive up). In performing BAEPs on unfamiliar equipment, care must be taken to select the correct montages so as to ensure that the displayed waveforms are of the correct polarity.

Common Mode-Rejection Ratio In an ideal op-amp, when the same signal is introduced at both input poles, there should be absolutely no signal at the output pole. In practice however, the op-amp’s ability to do this is imperfect. There is always a stray signal that is seen at the output when a common signal is introduced at the input. This was more

1/17/08

11:58 AM

Page 81

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt

Input Impedance The ACNS guidelines state that neurodiagnostic equipment’s amplification system should have an input impedance of at least 100 Mohm (1). Again, all modern commercially available NIOM machines meet this specification. High input impedance is important, as it ensures that most of the recorded signal voltage will be across the inputs of the amplifier. This is how it works: The equivalent circuit of the patient amplifier interface in its simplest form is shown in Figure 6.5. The impedance of the patient and input impedance of the amplifier are in series. Since resistors in series act as a voltage dividers, the voltage recorded across the patient and the voltage recorded across the amplifier are proportional to their respective impedances. Assume that the impedance of the electrodes on the patient is 1 Kohm and the input impedance across the

Patient

1KOhm

of a problem in the past, when vacuum tube amplifiers were used in neurodiagnostic equipment (this is where the terms G1 and G2 were introduced, as they referred to the metal grids seen in a vacuum tube amplifier). Today’s solid-state amplifiers are much more efficient; however, they are still not perfect. The efficiency of an amplifier’s ability to reject common input signals is expressed in the common mode-rejection ratio. According to ACNS guidelines, the common mode-rejection ratio of neurodiagnostic equipment should be at least 10,000:1 (or an 80-dB difference). For the purposes of this chapter, the specifics of how the common mode-rejection ratio is maximized are not important. A lot seems to be made of the common mode-rejection ratio; but for practical purposes, this ratio is within acceptable limits in all modern NIOM machines. Even the lowliest op-amp exceeds these common mode-rejection ratio specifications. It suffices to know that the higher the common mode-rejection ratio, the more efficient the amplification system is at discerning the differences between signals introduced at the input poles.

10MOhm

Husain 06

•

81

– Amp +

Voltage Divider Rule: The voltage recorded across the patient and across the amplifier is proportional to their impedance. Vamp is 100,000 greater than Vpatient. FIGURE 6.5 Simple representation of the patient, amplifier-equivalent circuit demonstrating the principle that as the impedance of both are in series, the voltage across each resistor in the circuit is proportional to its impedance.

amplifier is 100 MOhm. That means that the portion of the total signal voltage across the amplifier is 100,000 times greater than that across the patient. For all intents and purposes, all of the signal voltage is across the amplifier; only a minuscule amount (1 x 10–5) is across the patient electrode. Therefore the output signal is a true representation of the input signal.

Amplification The op-amp not only determines the morphology and polarity of its output signal, it also amplifies the signal so that it can be filtered, digitized, and displayed. The amplifier in an NIOM machine must be able to handle amplification of a wide variety of neurologic signals. EP signals are in the order of 0.1 – 5 µV, EEG signals are in the 10 to 100 µV range and EMG signals are in the millivolt range. So not only does an amplifier have to be sensitive enough to record very tiny signals, it has to be versatile enough to be able to amplify a wide range of signal types. It was not unusual in the not too distant past that in order to perform EPs, NCS, or EEG, separate pieces of equip-

82

1/17/08

•

11:58 AM

Page 82

SE C T I ON I : B a s i c P r i n c iple s

ment had to be brought into the operating room. Today, most NIOM equipment will perform all of these duties, not only separately, but also simultaneously.

Analog Filtering After amplification, the first of a series of processes that separate the neurologic signal from electrical and biologic noise occurs. This first step is processing the signal through a series of analog filters. A raw recorded signal is composed of a wide spectrum of component frequencies. Ideally, only that part of the raw signal that contains neurologic signal should be processed. The frequency response that is introduced in signal processing can be limited by the use of filters. Typically for NIOM procedures, low-cut and high-cut filters are utilized to delineate a signal with a limited band of frequencies that will be further processed. A low-cut filter (otherwise known as a highpass filter) will eliminate all component frequencies of the raw input signal lower than the set point. A high-cut (or low-pass) filter will eliminate all frequencies above its set point. Thus a bandpass of interest can be defined for each type of modality. Typically, for EEG, this is 1 to 70 Hz; for SEPs it is 30 to 3,000 Hz; for BAEPs, it is 10 to 3,000 Hz; and for EMGs, it is 10 to 32 kHz (Figure 6.6). An ideal bandpass filter, also known as a brick-wall filter, should eliminate all frequencies outside of its low- and high-frequency set points (Figure 6.7). However, real filters do

EMG BAEP SSEP EEG

0

10

100 1000 10,000 Frequency (Hz)

FIGURE 6.6 Representative bandpasses for electrically recorded biological activity, including EEG, SEP, BAEP, and EMG.

Amplitude (dB)

Husain 06

Bandpass Filter

Ideal

Actual

Log Frequency (Hz)

FIGURE 6.7 Amplitude-vs.-frequency plot showing characteristics of a bandpass filter. An ideal bandpass filter, or brick-wall filter, cuts off all frequencies outside of the selected endpoints of the filter. In reality, no analog filter behaves ideally; some portion of frequencies outside of the bandpass is allowed to pass.

not behave in this manner. In fact, all filters have a roll-off, where frequency components outside of the bandpass are allowed to pass. The sharper the slope of this roll-off, the more like an ideal filter the bandpass filter acts. The roll-off slope can be quantified in the number of decibels per octave that the signal degrades. The higher the absolute roll-off, the more like an ideal the filter behaves. In its simplest form, a low- or high-cut filter is an arrangement of resistors and capacitors or resistors and inductors. This simple filter with passive components is known as a first-order filter. It has a roll off of –6 dB per octave. This means that the signal strength will be cut in half each time the frequency doubles. Therefore, for a high-cut filter set at 3,000 Hz, about half the signal amplitude will be present at 6,000 Hz, one quarter at 12,000 Hz, etc. This is not an ideal situation. Analog filters with active components greatly improve the ability to exclude frequencies outside the bandpass. A second-order filter has a roll-off of –12 dB per octave. Third-order filters have a roll-off of –18 dB per octave, etc. The rolloff continues to improve with subsequent sophistication of filter design. The ACNS guidelines recommend that the low-cut filter have a roll-off of at least –12 dB per octave

1/17/08

11:58 AM

Page 83

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt

(second-order), and the high-cut filter have a roll-off of at least –24 dB per octave (fourthorder) (1). The physical properties of an analog filter introduces an error at its set point, or cutoff frequency. Instead of the entire signal passing at the cutoff frequency, the signal is degraded whereby a small portion of the signal is lost. In fact, there is a decrease of –3 dB in the signal at the cutoff frequency (Figure 6.8). This error is reflected in the ACNS guidelines where analog filter settings are discussed. For example, the ACNS EEG guidelines state that the low-cut filter should not be higher than 1 Hz (–3 dB), and the high-cut filter should not be lower than 70 Hz (–3 dB) (1). The –3dB in parenthesis recognizes that the signal at the set points of the bandpass will be degraded by –3 dB. When filter settings outside the recommended guidelines are used, distortions in the resultant waveforms can occur. Excessive use of low-cut filters (i.e., using a higher-than-recommended low-cut setting) will cause the resultant waveform to appear earlier. Excessive use of high-cut filters (i.e., using a lower-than-recommended high-cut setting) can cause a resultant waveform to appear later than normal. Excessive use of filtering or

Cut off frequency

Amplitude (dB)

Husain 06

–3 dB

Ideal

Actual

•

83

changing filter settings during an NIOM procedure can lead to confusion in waveform identification and can mask significant changes in neural function. Although waveforms can sometimes be “noisy,” it is always better to preserve the fidelity of your recorded waveforms by judicious use of filtering, than to risk masking their true nature through overuse of analog filters.

Artifact Rejection The next method of removing unwanted artifact from an NIOM EP recording is the use of automated artifact rejection. Artifact rejection causes signals over a certain voltage to be eliminated from subsequent processing. Therefore such signals will not be included in an EP average nor will they be displayed in a processed EEG trace. Artifact rejection is typically performed through a window discriminator, a device that compares the voltage of the incoming signal to a reference value (Figure 6.9A). If any portion of the input signal falls outside of the window, it is rejected (Figure 6.9B). It is important that the artifact rejection be user-controlled. Having the ability to adjust the size of the window in relation to the input signal is important in making the artifact-rejection feature efficient. Effective use of the window discriminator allows quicker acquisition of averaged responses. Therefore the user must be able to see both the raw input signal and the window in order to make this adjustment.

Central Computer System Analog-to-Digital Conversion

Log Frequency (Hz) FIGURE 6.8 Diagram showing a decrease of –3 dB in the signal at the cutoff frequency. See text for details.

The analog signal is then digitized for further processing, display, and storage. An analog-to-digital converter (ADC) is used for this process. There are several characteristics of the ADC that affects its ability to faithfully reproduce the analog signal. It must be understood, however, that the digitized waveform is a representation of the analog waveform, and

Husain 06

84

2/5/08

•

11:18 AM

Page 84

SE C T I ON I : B a s i c P r i n c iple s

A Input

Output

Input

Output

Upper Limit

Signal

Lower Limit

B Upper Limit

Signal

Lower Limit

FIGURE 6.9 Characteristics of a window discriminator used in artifact rejection. A. If the input signal falls within the upper and lower limits of the window discriminator, it is allowed to pass. B. If however, the input signal is over either the upper or lower limits of the window discriminator, it is rejected from subsequent processing.

inherent in the digitization process is loss of information. Modern NIOM machines do a very good job of ADC, with very good fidelity in the digitized signal. Once the signal has been digitized, it can be manipulated to further reduce noise through signal averaging and digital filtering. It can easily be displayed, trended, printed, and stored for future analysis. The discussion that follows delves into the characteristics of the ADC, explaining how it works and how these characteristics come together to ensure faithful reproduction of the analog signal.

Sampling Frequency, Dwell Time, and Horizontal Resolution Digitization of an analog signal involves several steps. First, a portion of the analog signal is isolated and held in memory. This analog signal may be an individual EP trial, a

NCS waveform, or an epoch of EEG or EMG data. The analog signal is then “sampled” at a certain distinct points along its time base. This sampling is done at a fixed frequency, therefore a fixed number of data points along the waveform are examined. The voltage of the analog waveform at each sample point is recorded as a discrete number. This is digitization. What results is a series of number pairs representing the time of each sample and the voltage of the analog signal at that time. This digitized data can then be used to plot the waveform on a screen, be used for signal averaging and digital filtering in processed EEG waveforms, and stored for future analysis. The fidelity of digital waveforms depends on how well the analog signal is sampled. The frequency with which the analog signal is sampled, or sampling frequency, must be high enough to ensure that enough data points are

Husain 06

1/17/08

11:58 AM

Page 85

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt

collected to faithfully represent the analog signal. The sampling frequency directly affects the horizontal resolution of the digitized waveform. The higher the sampling frequency, the more faithful the ADC representation of the analog signal will be. As seen in Figure 6.10, a sine wave is introduced into an ADC. In Figure 6.10A, the sampling frequency of the ADC is the same as the sine wave frequency. The resultant digitized waveform is a poor representation of the analog waveform; in this case, the digital waveform is a flat line. If the sampling frequency of the ADC is increased to twice the frequency of the analog sine wave (Figure 6.10B), the digitized waveform is a better but not a faithful representation of the analog signal. As shown in Figure 6.10C and D, increasing the sampling frequency to four and eight times the analog sine wave frequency greatly improves the ability to faithfully represent the analog waveform digitally. The theorem describing the minimum sampling frequency required for an ADC to faithfully represent an analog signal is known as the Nyquist theorem. It states that the sampling frequency of an ADC must be greater than twice that of the fastest-frequency component of a waveform. For example, the bandpass frequency of interest in a SEP waveform is 30 to 3,000 Hz. Therefore, to adequately represent the analog SEP waveform, the ADC must sample the waveform at greater than 6,000 Hz. That would mean that a sample would be taken every 0.00016 seconds, or every 0.16 ms, or 160 µs. This time between ADC samples is known as the dwell time. For EMG signals, where the upper end of the bandpass is in the range of 32,000 Hz, the sampling rate required would be in the 100-kHz range. Most modern NIOM machines have sampling frequencies in the megahertz range, so faithful ADC reproduction of neurologic signals is assured.

Bits and Vertical Resolution Just as the sampling frequency (and dwell time) determines the horizontal or time reso-

•

85

A: Sampling frequency = sine wave frequency Input A: Sampling frequency = sine wave frequency Input

A: Sampling frequency = sine wave frequency A: Sampling frequency = sine wave frequency Input Output Input Output Output

B: Sampling frequency = 2 x sine wave frequency Output Input B: Sampling frequency = 2 x sine wave frequency Input

B: Sampling frequency = 2 x sine wave frequency B: Sampling frequency = 2 x sine wave frequency Input Output Input Output Output Output

C: Sampling frequency = 4 x sine wave frequency C: Input Sampling frequency = 4 x sine wave frequency Input

C: Sampling frequency = 4 x sine wave frequency C: Input Sampling frequency = 4 x sine wave frequency Output Input Output Output Output

D: Sampling frequency = 8 x sine wave frequency Input D: Sampling frequency = 8 x sine wave frequency Input D: Sampling frequency = 8 x sine wave frequency D: Sampling frequency = 8 x sine wave frequency Input Output Input Output Output Output

FIGURE 6.10 The effect of various sampling frequencies on the output of an ADC. If the sampling frequency is the same as the frequency of the analog signal, the digitized output is a straight line. As the sampling frequency of the ADC is increased, the fidelity of the digitized signal improves.

Husain 06

86

1/17/08

•

11:58 AM

Page 86

SE C T I ON I : B a s i c P r i n c iple s

lution of an ADC, the number of bits available to the ADC determines the vertical, or amplitude, resolution of the digitized waveform. As discussed previously, the ADC converts the voltage of the analog waveform to discrete digital numeric data at each sample point. The number of vertical data points available determines how faithful the amplitude data are translated by the ADC. The number of vertical data points available for the full scale voltage of the analog waveform is expressed in bits. The number of vertical data points available per bit is expressed by the equation 2n, where n 5 the number of bits available to the ADC. Therefore an eight-bit ADC would have 28 or 256 data points of vertical data resolution. Thus, when displaying a waveform that has a peak-to-peak voltage of 10 µV, an eight-bit ADC would in theory be able to resolve and display voltage changes of 0.04 µV. This is an ideal situation; in the real world, there is an error factor involved. A 12bit ADC would have 212 or 4,095 data points of vertical resolution, meaning that in theory it could resolve voltage changes of 0.002 µV for the same 10-µV signal. NIOM machines with inadequate vertical resolution were a factor back in the days when laboratories assembled their own neurodiagnostic equipment. However, most modern NIOM machines have at least 12-bit ADC.

The EP Waveform, Signal-to-Noise Ratio, and Signal Averaging Signal-to-Noise Ratio EP signals are tiny. When recorded from the scalp, they are buried in a myriad of electrical and biologic noise often several magnitudes larger than the signal of interest. EP signals are in the single-digit microvolt range. Ambient electrical (60 Hz) noise is 120 V at the source, but can be in the millivolt or volt range at or near the amplifier. EEG activity is in the 10-µV to 100-µV range; EMG and EKG activity is in the millivolt range. Bandpass filtering does not get rid of all this electrical and

biologic noise. As seen in Figure 6.9, there is a lot of overlap in the component frequencies of EP, EEG, EMG, and 60-Hz electrical activity. A comparison of the amplitude of the EP signal to the amplitude of the background noise can be made. The signal-to-noise ratio (SNR) is just that, SNR 5 Vsignal/Vnoise. EP waveforms buried in background noise have low SNRs. EPs relatively free from background noise have greater SNRs. Filtering by itself, as discussed above, is limited in its ability to increase a waveform’s SNR. Other techniques, the most important being signal averaging, are used to increase the ability to tease the EP out of background noise.

Signal Averaging Signal averaging utilizes the principle that EPs are time-locked to their stimuli and all other background activity is random. If a sequence of EP trials is averaged together, the time-locked potential will be present in all trials while the random background activity will eventually cancel itself out. The caveat here is that all background activity must be random. This becomes a problem with 60-Hz noise. If the EP stimulus is synchronized to the 60-Hz signal, or some harmonic of 60 Hz, the 60-Hz signal will be included in the averaged waveform. Therefore it is imperative, in performing EP studies, that the stimulator be set at a repetition rate that is not a factor of 60 or its harmonics. In Figure 6.11, an idealized EP waveform is presented with representations of five individual EP trials. Each EP trial contains the time-locked EP waveform and random background noise. As represented in Figure 6.12, a series of subsequent EP trials are averaged together. As the number of trials in the average increases, the EP waveform emerges from the background noise. Increasing the number of trials in the average further improves the signal until the background activity is negligible and the EP waveform is clear. As subsequent EP trials are averaged, the SNR improves. The improvement in the SNR

Husain 06

1/17/08

11:58 AM

Page 87

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt Ideal waveform

Trial 1

Trial 2

Trial 3

Trial 4

Trial 5

FIGURE 6.11 An ideal EP waveform as well as five representative individual trials. Each of the trials contains the idealized EP waveform as well as background biological and extraneous electrical noise. Signal averaging is used to separate the timelocked EP waveform from the background noise.

•

87

waveform that contains the average of 16 sweeps would have a four time improvement in its SNR; 64 sweeps would improve the SNR by a factor of 8, or 256 sweeps by a factor of 16, or 1,024 sweeps by a factor of 32, etc. Early signal averagers and clinical EP machines had selections for number of averages that were squares of whole numbers for the simple reason that it was easy to calculate the SNR improvement. Here is a sample problem that illustrates how SNRs are used in the laboratory: A 1-µV EP signal is buried in 5 µV of background noise. An appropriate SNR for this study is 5. How many averages would be necessary to accomplish this? 1. First calculate the starting SNR: SNR 5 1 µV / 5 µV 5 0.2 2. The target SNR is 5, which is 25 times greater than the present SNR. 3. The SNR must be improved by 25, which means that 252 or 625 trials will be needed to achieve the desired goal.

Ideal EP waveform

Digital Filtering and Smoothing # trials averaged

FIGURE 6.12 An idealized EP waveform as well as averaged waveforms with increasing number of trials. As the number of trials increases, random background noise is diminished until a true representation of the idealized waveform is obtained.

can be expressed mathematically, where improvement is equal to the square root of the number of trials averaged. For example, an EP

The digitized waveform can undergo digital filtering to eliminate the roll-off characteristic of analog filters. Digital filters act as the ideal brick-wall filter with no roll-off. In one common form of digital filtering, the digitized waveform undergoes a Fourier transform, where the amplitude of the waveform in individual frequency bands are determined. Waveform components in frequency bands outside of the selected bandpass are eliminated, and the digital waveform is reassembled from the remaining data. Digital filters do not suffer from the phase shifting inherent in analog filters, therefore there is no temporal distortion of the resultant digitally filtered waveform. Most EP equipment offers some form of smoothing involving software algorithms that appear to increase the SNR by eliminating a portion of high-frequency background activ-

Husain 06

88

1/17/08

•

11:58 AM

Page 88

SE C T I ON I : B a s i c P r i n c iple s

ity. These algorithms vary; in their simplest form, however, they take a weighted average of the data points in the waveform and fit a curve to the data points that best represent this weighted average. In this way, stray highfrequency data are eliminated. Just as with other forms of digital filtering, phase shifts do not occur, but the amplitude and the morphology of the real waveforms do change. Digital filtering and smoothing, just like analog filtering, should be used carefully and their effects on the waveform understood.

Data Display, Trending, and Storage After amplification, filtering, digitization, and signal processing, the final process of the NIOM machine is to display the resultant data in a manner that allows changes to be identified and tracked efficiently. NIOM machines all have the ability to display and trend data. In general, one should be able to see the live data, compare it to data previously recorded, set baseline data, be able to trend visual data to numerical data, and place timelocked comments to data sets. For EEG and EMG data, a machine must have a clear and easy method of reviewing data on the fly. At best, the complete raw EEG and EMG data should be stored and a mechanism for reviewing data on the fly should be available. At the very least, several minutes of EEG and EMG data should be held in a buffer in order for transient events to be reviewed or replayed. The NIOM machine should be capable of flagging epochs of EEG and EMG data, such as baseline data or data at critical periods. These flagged data should be easily accessible for comparison to current “live” data. A splitscreen method of data presentation with stored data on one side of the screen and live data on the other is an excellent method of being able to compare EEG and EMG data of interest. It is imperative that the NIOM machine be able to transmit data online and in real time to the supervising neurologist or neuro-

physiologist. The advent of wireless Internet connections makes this task easier than in previous years. One concern of data transmission is ensuring that that there is enough bandwidth to transmit raw EEG and EMG data accurately. Data transmission must adhere to Health Information Portability and Accountability Act (HIPAA) guidelines to protect patient confidentiality. After completion of the case, a method for storing the data for review and archiving should be available. Although most personal computer–based machines have software for writing files to CDs or DVDs, it is much more convenient for NIOM software to be able to transfer data instead of using another standalone program. In addition, there should be a method of flagging which data have been reviewed and which have been archived to CDs or DVDs. Previous case data should be removed from the NIOM machine as soon as practible and stored with the patient’s permanent record. This is important for two reasons. First, it ensures that there is sufficient room on the NIOM machine’s hard drive for future studies. Second, it ensures that HIPAAprotected patient data are not residing on a hard drive that may be accessed by unauthorized persons.

EP, NCS, and tceMEP Data EP displays should have the several features. The current average should be clearly displayed. Baseline waveforms for each trace in the montage should be easily set and prominently identified. It is a good idea to have a display that shows the current average and the baseline in a manner that allows the traces to be visually compared. A trend of previous averaged traces is helpful in visually determining changes in waveform morphology. It is also important that numeric data be displayed in such a way as to easily determine if significant changes from baseline have occurred. Time-locked comments linked to waveforms and numeric data reflect the function of the nervous system at the various stages of surgery

Husain 06

1/17/08

11:58 AM

Page 89

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt

and link changes in waveform morphology to what has occurred in the operative field. A clear representation of the various stimulating and recording parameters should be available. Stimulating parameters should include rate, intensity, and pulse width. Recording parameters should include number of averages, gain, and bandpass information. Data should be stored in such a way that individual averages, waveform trends, and numeric data can all be retrieved for any or all portions of the surgery.

EEG Data Intraoperative EEG data consists of the raw signal and processed EEG data such as compressed spectral array (CSA) and density spectral array (DSA). The raw EEG data should be available at any time. Digital EEG technology allows the display montage to be changed at any time during the procedure to better localize changes. Unlike paper EEG, montages can be changed on the fly without losing the ability to go back and remontage stored data. The entire EEG recording should be stored for review. It is also helpful to be able to split the screen during recording and to compare the live EEG to the previously recorded EEG. Event markers with comments should be available, allowing the technologist to quickly search the record for significant events as well as to correlate changes in the EEG to events in the operative field. The advent of CSA and DSA EEG allows the technologist to quickly identify major trends in frequency shifts and loss of amplitude or power. They are a good generalization of what is happening to the EEG but cannot replace vigilant interpretation of raw EEG data. They can miss subtle changes that may be clinically significant. All changes in processed EEG data have to be correlated with the raw EEG.

EMG Data The NIOM machine should be able to display and capture multiple channels of free-

•

89

running EMG data. At best, the live EMG record for the entire case should be stored, as is the case with EEG. If this is not possible, time- and event-marked epochs of EMG data should be stored for analysis. It is useful to have several seconds or minutes of EMG data in a buffer so as to be able to review the EMG data and store relevant epochs of activity. The ability to output audio EMG activity is helpful in providing feedback to the surgeon when he or she is irritating neural structures. Care must be taken in using audio output in the presence of electrocautery. Most NIOM machines have a method of muting the audio output when an electrocautery signal is detected. Many NIOM machines have an autocapture function using a window discriminator. When the EMG signal is of greater amplitude than the threshold of the window, timestamped EMG activity is captured on the screen and stored. This is very helpful when dealing with frequent EMG activity in multiple channels.

Summary This chapter reviews the requirements of an NIOM machine to adequately record data in the operating room. The ACNS guidelines provide recommendations for the minimum features that must be present in these machines. Most commercially available machines meet these guidelines but differ in other features. The various features of most commercially available NIOM machines are compared in Table 6.2. It is recognized that NIOM machines are constantly being updated, and further changes will occur with time in each of these machines. It is hoped that this chapter will help the reader to decipher the features of various NIOM machines and guide him or her in finding a system that will meet most needs. Once one understands what each part of the NIOM machine does and the reasons why certain features are important, one should be able to uti-

32 Y Y Y Y Y Y Y HD, CD, DVD

# Channels

EP

BAEP

VEP

EMG

EEG

Integrated tceMEP

Multimodality

Storage

N/A > 100 dB < 2 uV p-p N/A N/A

Input impedance

CMRR

Noise level

Sensitivity

Time base

N/A N/A N/A 50/60 Hz

Order/rolloff

LFF

HFF

Notch

Filters

N/A

ADC

50/60 Hz

30–10 KHz

0.5–100 Hz

2nd/12dB/octave

1–1000 msec/div

0.01 µV–10 mV/div

< 4uV p-p

> 95 dB

> 100 Mohm

16 bit, 25.6 KHz

HD, CD, DVD

Y

Y

Y

50/60 Hz

10 Hz–3 KHz

0.08 Hz–3 KHz

1st or 2nd/6 or 12 dB/octave

1–6000 msec/div

0.05 µV–50 mV/div

< 3uV p-p

>106 dB

> 100 Mohm

16 bit, 5us/channel

HD. CD, DVD

Y

N

Y

Y

Y

Y

Y

50/60 Hz

100–3000 Hz

0.2–500 Hz

1st or 2nd/6 or 12 dB/octave

N/A

10 µV–100 mV scale

0.7 µV RMS

110 dB

> 1000 MOhm

N/A

HD, CD, DVD

Y

N

Y

Y

Y

Y

Y

16

Viasys Endeavor CR

50/60 Hz

30–15 KHz

0.1–500 Hz

N/A

0.5–500 msec/div

0.1 µV–5 mV/div

110dB

>100 MOhm

16 bit, 60KHz

HD, CD, DVD

Y

N

Y

Y

Y

Y

Y

16

XLTEK Protektor

11:58 AM

Y

Y

Y

Y

16 or 32

Nihon Kohden Neuromaster

•

16 (32 Cascade Elite)

Cadwell Cascade

1/17/08

Amplifier

Axon Eclipse

90

Manufacturer Model

TABLE 6.2 Comparison of the Commonly Used NIOM Machines Currently Commercially Available

Husain 06 Page 90

SE C T I ON I : B a s i c P r i n c iple s

0.01–100 Hz Y 0 –100 mA (0–400 V) 0.1–500 msec Y

Rate

Interleave

Intensity

Duration

External Trigger

N/A N/A

Intensity

Polarity

N/A: Not available.

Tone/burst/click

Signal Type Cond/rare/alt

–10–95 dB nHL

Click

N/A

0.05–1 msec

0–100 mA

Y

0.5–90 Hz

16

Cadwell Cascade

Cond/rare/alt

0–135 dB SPL

Tone/ burst/click

Y

0.05–1 msec

0–100 mA

Y

0.1–50 Hz

6

Nihon Kohden Neuromaster

Cond/rare/alt

–21–103 db nHL

Click

Y

0.01–1 msec

0 –100 mA (0–400 V)

Y

0.1–100 Hz

4 high level, 1 low level

Viasys Endeavor CR

Cond/rare/alt

0–105 dB nHL

Click/pip/tone/noise

Y

0.05–1 msec

0–100mA

Y

0.1–100 Hz

16

XLTEK Protektor

11:58 AM

Auditory

16

Axon Eclipse

1/17/08

Number

Estimated

Manufacturer Model

Husain 06 Page 91

CH A P T E R 6: A Buye r’s Guid e to Monitoring Equipme nt

• 91

Husain 06

92

1/17/08

•

11:58 AM

Page 92

SE C T I ON I : B a s i c P r i n c iple s

lize all the features of the equipment to their fullest, making NIOM more effective.

REFERENCES 1. American Clinical Neurophysiological Society. Guidelines 1–10. Am J Electroneurodiagn Technol 2006;46(3):198–305.

2. Moller AR, Janetta PJ. Preservation of facial function during removal of acoustic neuromas: Use of monopolar constant-voltage stimulation and EMG. J Neurosurg 1984;61: 757–760.

Husain 07

1/17/08

11:58 AM

II

Page 93

Clinical Methods

Husain 07

1/17/08

11:58 AM

Page 94

Husain 07

1/17/08

11:58 AM

7

Page 95

Vertebral Column Surgery

David B. MacDonald Mohammad Al-Enazi Zayed Al-Zayed

S

urgery on the vertebral column risks spinal cord, nerve root, brachial plexus, and peripheral nerve injury. Although the overall incidence of major neurologic complications is only about 1%, the devastation of paraplegia strongly motivates efforts to monitor spinal cord integrity during these procedures (1). Consequently spine surgery was one of the first and remains one of the most frequent indications for neurophysiologic intraoperative monitoring (NIOM). Two general characteristics tend to make these procedures ideal for monitoring: (a) most patients have monitorable evoked potentials (EPs) because they are neurologically intact and (b) the most commonly encountered injury mechanisms (ischemia, compression, traction) are reversible when detected quickly. Nevertheless, a few patients having antecedent neurologic pathology degrading their EPs can be challenging or even impossible to monitor. In addition, some injury mechanisms, such as contusion, may not be reversible, although it may be possible to minimize their effects if the cause is rapidly identified and corrected. This chapter outlines the basic anatomy, pathology, clinical features, and surgery of common vertebral column disorders and then details NIOM techniques, interpretation, and technical issues. It emphasizes spinal cord

monitoring; nerve root protection techniques that may also be relevant are addressed elsewhere in this book.

BASIC ANATOMY, PATHOLOGY, CLINICAL FEATURES, AND SURGERY

Vertebral Column Anatomy The vertebral column consists of seven cervical, twelve thoracic, and five lumbar vertebrae, their interposed fibrocartilaginous discs, and connective tissues. It is normally straight in the coronal plane, but in the saggital plane it curves gently forward in the neck, backward in the thorax (normal kyphosis), and forward again in the lumbar spine. Each cylindrical vertebral body has a left and right pedicle projecting posterolaterally to the lateral transverse processes that articulate with neighboring transverse processes at facet joints. The left and right laminae project posteromedially from the transverse processes to join together as the midline spinous process. Longitudinal ligaments run down the ventral and dorsal surfaces of the vertebral bodies and discs; interspinous ligaments connect adjacent spinous processes, and the ligamentum flavum connects adjacent laminae. 95

Husain 07

96

1/17/08

•

11:58 AM

Page 96

SE C T I ON I I : C l i n i c a l M e thod s

Spinal Cord Anatomy The vertical space between the vertebrae, their processes, and the ligaments forms the spinal canal containing the spinal cord, which normally ends at about the L1 vertebral level. Cervical and thoracic spinal nerve roots exit the spinal canal laterally through the foramina formed between adjacent pedicles. Lumbosacral roots pass downward as the cauda equina to exit through foramina below the spinal cord. The corticospinal tracts important for voluntary movement descend from the cortex, decussate (cross the midline) in the caudal medulla, and then travel down the contralateral dorsolateral cord. Some of their axons form excitatory synapses on alpha motor neurons in the anterior horn gray matter, mostly contralateral to the originating hemisphere. A few corticospinal fibers descend without decussation through the ipsilateral ventral cord but are normally of minor importance. The axons of alpha motor neurons from each spinal cord segment coalesce to form motor roots (e.g., C5, L4), which eventually innervate muscles. Cervical and lumbosacral roots undergo complex interdigitation within the brachial and lumbosacral plexi, which rebranch into the various peripheral nerves supplying limb muscles. Consequently there is radicular overlap and anatomic variation of muscle innervation. Intraoperative monitoring of motor EPs (MEPs) can assess the corticospinal system from brain to muscle. The dorsal columns contain sensory axons for position and vibration sense. These very long axons ascend the ipsilateral dorsal column to the gracile and cuneate nuclei at the cervicomedullary junction, where they finally synapse. The axons of the second-order sensory neurons then decussate as the internal arcuate fibers of the caudal medulla before ascending to the contralateral thalamus, where they synapse on third-order sensory neurons. These then project to somatosensory cortex. Intraoperative monitoring of

somatosensory EPs (SEPs) can assess this system from peripheral nerve to brain. The spinothalamic system conveying pain and temperature sensation and many other descending, ascending, and intrinsic spinal cord systems are unassessed by MEP/SEP techniques. F-wave, H-reflex, and bulbocavernosus reflex testing can provide information about segmental cord activity that may indirectly help assess long tract functions.

Spinal Cord Blood Supply Radicular arteries from the aorta and its major branches in the neck and pelvis enter the spinal canal through the intervertebral foramina. With considerable variation, a few of these arteries branch to feed the longitudinal anterior spinal artery running down the anterior median fissure of the spinal cord and the left and right posterior spinal arteries running down the dorsal spinal cord. The anterior spinal artery supplies the ventral two-thirds of the cord including the anterior horns, central gray matter, ventral portions of the dorsal horns, and corticospinal tracts. The posterior spinal arteries supply the dorsal columns and superficial dorsal horn gray matter. A pial plexus of interconnecting arteries supplies the outer rim of the cord.

Vertebral Column Pathology Scoliosis Scoliosis consists of abnormal vertebral column curvature in the coronal plane. Cobb angle measured in spinal x-rays is an estimate of the degree of deformity. There is also a rotational deformity and there may be excessive kyphosis (kyphoscolsiosis). Following the shortest vertical path, the spinal cord adopts an ectopic and rotated position toward the concave side of the deformed spinal canal. Most often scoliosis is idiopathic. Congenital scoliosis is also common, consisting of vertebral column malformation associated with hemivertebrae and other vertebral anomalies. Vertebral dysplasia due to neurofi-

Husain 07

1/17/08

11:58 AM

Page 97

CH A P T ER 7: Ve r te bral Column Surge r y

bromatosis or achondroplasia is another relatively common cause. Connective tissue disorders such as Turner’s syndrome are less frequent causes. Several neurologic disorders are associated with scoliosis, including cerebral palsy, Arnold-Chiari malformation, Freidreich’s ataxia, spinal cord tumor, neural tube defects, tethered cord, and neuromuscular disease. These are grouped together under the term “neuromuscular scoliosis.” Horizontal gaze palsy and progressive scoliosis (HGPPS) is a rare autosomal recessive disorder consisting of an inability to turn the eyes to either side along with progressive scoliosis (2). Although more frequent in regions where consanguinity is common (it comprises about 3% of the authors’ scoliosis surgery series in Riyadh, Saudi Arabia), it has also been reported in Europe, Japan, and North America. The corticospinal and dorsal column systems are uncrossed (nondecussation) and NIOM techniques must accordingly be modified for accurate results (3).

Other Vertebral Column Disorders Some other surgical spinal disorders that may require NIOM include cervical spondylosis, spinal fractures, vertebral column tumors, and ossification of the posterior longitudinal ligament (OPLL). Spondylosis is a degenerative process consisting of bony spur formation on the lips of adjacent vertebrae along with disc degeneration and prolapse. Spinal fractures normally result from substantial trauma unless there is a predisposing bone disease. Primary or metastatic vertebral column tumors may involve the vertebral bodies, their bony processes, and adjacent tissues. OPLL affects mostly Japanese patients and causes cervical spinal canal narrowing due to calcification and ossification of the posterior spinal ligament.

Clinical Features Scoliosis Idiopathic scoliosis usually presents as spinal deformity of a child or teenager

•

97

noticed by the family or on routine medical examination. There may be associated back pain but usually no neurologic deficits. Only rarely, the disorder can be severe enough at presentation to have caused spinal cord compression with varying degrees of myelopathy. Severe thoracic scoliosis with rib cage deformity that impairs the movement of the chest wall may cause restrictive lung disease, but this is similarly rare with idiopathic scoliosis. Patients with congenital or dysplastic causes tend to present at an earlier age and are more likely to have preoperative neurologic or pulmonary complications. Patients with neuromuscular scoliosis have neurologic pathology that may reduce the effectiveness of NIOM. In planning for and anticipating the results of monitoring, one should consider the nature and severity of the underlying neurologic disease. Cerebral disease such as cerebral palsy may interfere with MEP and cortical SEP monitoring. Posterior fossa disorders such as ArnoldChiari malformation may cause EP abnormalities. Spinal cord disorders such as syringomyelia, diastamatomyelia, tethered cord syndrome, and spinal cord tumor, compression, or trauma may cause varying degrees of sensorimotor impairment and/or EP degradation that occasionally seems out of proportion to any clinical deficit. The chronic corticospinal and dorsal column pathway degeneration of Freidreich’s ataxia can prevent effective NIOM. Pure motor system diseases like spinal muscular atrophy, polio, and muscular dystrophy interfere with muscle MEPs but do not affect SEPs. Peripheral neuropathies may interfere with both modalities. HGPPS patients have no clinical sensorimotor deficits and, with techniques appropriately modified for their nondecussation, have easily monitored potentials (3). Because the patients naturally compensate for the gaze palsy by turning their heads, they, their families, and their physicians may not notice the abnormality unless it is specifically sought.

Husain 07

98

1/17/08

•

11:58 AM

Page 98

SE C T I ON I I : C l i n i c a l M e thod s

Other Vertebral Column Disorders Cervical spondylosis presents in older patients with neck pain, radiculopathy, and sometimes myelopathy with segmental and long tract sensorimotor deficits. OPLL generally presents with cervical myelopathy. Vertebral column tumors cause pain, radiculopathy, and myelopathy when cord compression is present. Spinal fractures may or may not be associated with acute traumatic myelopathy, sometimes paraplegia or quadriplegia. Again, the nature and severity of antecedent neurologic deficits should be incorporated into the NIOM strategy and expectations.

Vertebral Column Surgeries Diverse surgeries are performed on the spine. There must be a tangible risk of injury to justify spinal cord monitoring. Scoliosis surgeries, usually posterior spinal fusion and instrumentation (PSF), anterior spinal release, or both are the most frequent indications. Some others include anterior cervical discectomy and fusion, decompressive laminectomy for spondylotic myelopathy or ossification of the posterior longitudinal ligament, reduction and fixation of spinal fractures, and spinal tumor resections.

Scoliosis Posterior spinal fusion and instrumentation for scoliosis correction basically involves straightening with bilateral metallic rods anchored to vertebral processes of the prone patient’s exposed spine. The anchors consist of pedicle screws and sublaminar hooks or wires. A properly directed pedicle screw passes through the pedicle’s center into the vertebral body without breaking through the pedicle wall into the spinal canal or foramina. Sublaminar hooks and wires enter the spinal canal. To straighten the spine, anchor pairs along the rod are strategically forced apart (distracted) or pushed together (compressed) along with any necessary rod rotation before

rigidly fixing attachments. Laminae are then decorticated and cancellous bone fragments spread over them to promote subsequent osseous fusion. Anterior spinal release may be necessary before PSF for patients with rigid and congenital scoliosis who have vertebral malformations. The procedure is performed endoscopically or through a thoracotomy with the patient in a lateral position. Intervertebral discs and malformed vertebrae are resected; associated radicular arteries are sacrificed. The PSF may immediately follow or take place in a subsequent surgery. Sometimes anterior fixation using metal screws, rods, or plates is performed without PSF. The large forces applied to the spine can injure the spinal cord through compression, traction, or ischemia, which may be potentiated by the use of controlled hypotension (60 mmHg mean arterial pressure) intended to reduce blood loss. The insertion of sublaminar hooks or wires can cause cord contusion, compression, or ischemia. This is more likely with insertions on the deformity’s concave side, where the spinal cord is located, or where the spinal canal is narrow. Misdirected pedicle screws can perforate through the pedicle wall and cause adjacent nerve root injury. Misdirected cervical or thoracic screws can compromise the spinal cord. During anterior release, acute spinal deformity following vertebrectomy or interference with critical radicular blood supply can cause compressive or ischemic cord injury.

Other Procedures Anterior cervical discectomy and fusion is performed through neck dissection with the patient supine. Involved discs are exposed and resected along with any posterior bony spurs. Bone grafts to produce subsequent osseous fusion are tapped into the evacuated disc spaces. A metallic plate is screwed onto the anterior surfaces of adjacent vertebrae for fixation. Spinal cord complications can result from contusion, compression, or ischemia.

Husain 07

1/17/08

11:58 AM

Page 99

CH A P T ER 7: Ve r te bral Column Surge r y

Posterior decompression is achieved through laminectomies. Various types of instrumentation are used to reduce and fix spinal fractures and subluxations. Spinal tumor surgery involves resection of tumor and associated bony structures through posterior, anterior, or combined approaches. This may include vertebral bodies, sometimes at multiple contiguous levels. Bone grafts, plates, and other devices are used for fixation and subsequent fusion. Any of these procedures can injure the spinal cord or nerve roots. A special consideration for cervical surgeries is that positioning the patient with neck extension or flexion can cause cord injury through cord compression, particularly when there is already compressive cervical myelopathy.

Monitoring Techniques Because motor deficit—paraplegia or quadriplegia—is the primary concern, it is illogical to rely on SEPs mediated through the dorsal columns. However, SEPs were the first applied modality because MEP technology was unavailable. Their application was based on the premise that spinal cord compromise threatening paralysis might affect the full transverse extent of the cord, thereby causing SEP deterioration due to dorsal column conduction block. Intervention to relieve cord compromise might then be undertaken to restore potentials and hopefully prevent injury. Indeed, SEP monitoring halves the risk of intraoperative cord damage and therefore remains a standard NIOM component today (1). However, due to the anatomic separation and distinct blood supplies of the corticospinal and dorsal column systems, discrete compromise of one or the other still occurs. Consequently, occasional motor compromise without SEP warning or SEP deterioration without motor compromise is unavoidable, and SEP monitoring alone is no longer sufficient now that MEPs are available. Conversely, MEP monitoring cannot predict dorsal column

•

99

integrity; therefore the two modalities should be combined (4).

Somatic Evoked Potentials No universally standardized SEP monitoring approach exists. The International Federation of Clinical Neurophysiology (IFCN) published guidelines in 1999 (5). The American Clinical Neurophysiology Society (ACNS) guidelines, last published in 1994, are undergoing revision, and an American Society of Neurophysiologic Monitoring (ASNM) position statement was published in 2005 (6,7). Upper limb SEPs are advised for cervical spinal cord monitoring, although the legs are also at risk. Lower limb SEPs are advised for thoracic spinal cord monitoring, but two guidelines point out the value of including upper limb SEP systemic controls (5,7). It is good practice to monitor all four limbs for both levels. Median nerve stimulation is generally recommended for upper limb SEPs, but the ulnar nerve is suggested instead when surgery risks C7-C8 cord injury that might be missed by median SEPs. The tibial nerve at the ankle is recommended for lower limb SEPs, with the peroneal nerve at the knee as an alternate site. To avoid the chance that bilateral simultaneous lower limb testing might mask unilateral cord compromise, the testing of each side separately, using alternating stimulation and asynchronous parallel averaging, is unanimously recommended. All three guidelines recommend scalp recordings of cortical potentials. They differ somewhat on montages, but all recommended derivations assume normal sensory decussation; derivations for detecting and monitoring patients with nondecussation are not addressed. Subcortical potentials and caudal or peripheral SEP controls are also recommended. None specifically base montage recommendations on derivation signal-to-noise ratio (SNR) and its profound effect on reproducibility and the rapidity of feedback. The

Husain 07

100

1/17/08

•

11:58 AM

Page 100

SE C T I ON I I : C l i n i c a l Me thod s

ASNM guidelines do point out the importance of rapid surgical feedback and suggest 300 to 500 sweep averages, but ACNS guidelines recommend time-consuming averages of 500 to 2,000 sweeps, and IFCN guidelines do not address this issue. IFCN and ACNS guidelines suggest the additional use of invasive spinal SEP monitoring.

Signal- to-Noise Ratio and Feedback Rapidity By enhancing surgical correlation, rapid feedback helps determine the likely cause of a pathologic EP decrement, thereby guiding intervention. Furthermore, the likelihood of successful intervention probably increases with early detection. Surgical feedback occurs at intervals determined by SEP averaging time. For reliable interpretation, SEPs must be averaged to reproducibility, defined as less than 20% to 30% random trial-to-trial amplitude variation and convincing waveform superimposition in successive trials. Since SNR has a powerful nonlinear inverse relationship to the necessary averaging time, even modest gains substantially accelerate feedback, so that one should strive to maximize SNR and reject low-SNR techniques (8).

Cortical Potentials Scalp cortical SEPs are noninvasive, relatively easy to obtain, and applicable to surgery at any neuraxis level. Consequently they are almost universally employed; only a few centers emphasizing invasive spinal SEPs omit them. Given this widespread reliance, methods to optimize cortical SEP SNRs should be used. Anesthesia is important in this regard. Traditional balanced anesthesia with halogenated gases and nitrous oxide produces marked dose-related depression of cortical SEPs and should not be used. Satisfactory results may be possible using less than 0.5 minimum alveolar concentration (MAC) halogenated gases and opioids without nitrous oxide. However, it currently appears that intravenous anesthesia such as propofol/

opioid infusion is optimal because of less dose-related amplitude suppression, thus enhancing SNRs (5). Cortical SEPs should be further optimized by derivation selection. Standard laboratory derivations are not necessarily appropriate for NIOM because the goals are different. In the laboratory, valid comparison to normative data requires standardized derivations; the primary interpretive criterion is latency, and SNR is not critical because there is time to average many sweeps. In the operating room, derivations need not be standardized because patients are their own controls, amplitude change is the primary criterion and SNR is critical. Therefore the highest SNR derivation should be selected for monitoring, and this can vary between patients and sides (8). The selection process is straightforward for upper limb SEPs, more complex for lower limb SEPs, and in either case should include decussation assessment.

Upper Limb Derivations Following median or ulnar nerve stimulation in a patient with normal sensory decussation, a negative ‘N20’ peak generated by the contralateral primary sensory cortex for the hand occurs at the contralateral postcentral scalp and a positive ‘P22’ potential arises over the frontal scalp. With nondecussation, the N20 is ipsilateral instead because the ipsilateral hemisphere generates it. Stimulating one side while simultaneously recording from CPc and CPi (midway between 10 and 20 central and parietal sites, contralateral and ipsilateral to the stimulated nerve) assesses decussation (3) (Figure 7.1). Thus, input 1 of the monitoring derivation should be CPc for normal decussation or CPi for nondecussation. A frontal input 2 (FPz or Fz) boosts signal amplitude because the inverted P22 adds to the N20. This makes CPc-FPz generally satisfactory for NIOM, and it is tempting to use it routinely. However, nondecussation produces an inverted P22 in CPc-FPz that will be inaccurately mistaken for

Husain 07

1/17/08

11:58 AM

Page 101

CH A P T ER 7: Ve r te bral Column Surge r y

•

1 01

FIGURE 7.1 Median SEP decussation assessment. Mastoid reference; FPz can be used instead. RBr and LBr, right and left Brachial peripheral SEP traces. A. Normal decussation B. Nondecussation.

a small N20 unless decussation is assessed (Figure 7.2). Furthermore, frontal sites contain more anesthetic electroencephalographic (EEG) noise than centroparietal sites and, by overwhelming the amplitude advantage, this can reduce SNRs and slow feedback (8) (Figure 7.3). Because there is often substantially lower noise with CPc-CPi or CPc-CPz, these can thereby boost SNRs, thus accelerating feedback. The short interelectrode distance of CPc-CPz can further decrease noise, but in-phase cancellation when the N20 field reaches the midline may reduce signal amplitude. When this does not occur, it is often optimal. In summary, the process for selecting the optimal upper limb cortical SEP monitoring derivation is to (a) assess decussation; (b) compare CPc-FPz, CPc-CPi, and CPc-CPz with normal decussation or CPi-FPz, CPi-CPc and CPi-CPz with nondecussation; and (c) use

FIGURE 7.2 Nondecussation and CPc-FPz. The median N20 arises over the hemisphere ipsilateral to the stimulated nerve. An inverted P22 (Inv. P22) from FPz masquerades as a small N20 in CPc-FPz, so that this derivation alone will miss nondecussation.

the derivation showing fastest reproducibility (highest SNR).

Lower Limb Derivations Tibial nerve stimulation produces a positive ‘P37’ scalp potential generated by leg sensory cortex of the contralateral hemisphere with normal decussation and of the ipsilateral hemisphere with nondecussation. This potential’s scalp field is highly variable (8–10). It is commonly maximal at CPz but may be maximal at Cz or Pz instead. With normal decussation, its field spreads over the scalp ipsilateral to the stimulated nerve while a simultaneous negative ‘N37’ potential usually arises over the contralateral scalp. The ipsilateral P37 field is sometimes so prominent that the maximum occurs at iCPi (intermediate centroparietal site CP1 or CP2, ipsilateral to the stimulated nerve) or even CPi rather than at the midline. In this case, the P37 may be small or virtually

FIGURE 7.3 Frontal anesthetic EEG patterns. Higher EEG noise content in FPz derivations than centroparietal derivations reduces SEP signal-tonoise ratio.

Husain 07

102

1/17/08

•

11:58 AM

Page 102

SE C T I ON I I : C l i n i c a l Me thod s

absent at the midline. Occasionally, when the P37 maximum is at Cz, the N37 is ectopic at Pz. Furthermore, P37/N37 topography is most often different on each side. Finally, nondecussation reverses P37/N37 lateralization and cannot be detected by CPz-FPz alone (3). Simultaneously recording from CP4, CPz, and CP3 while stimulating one side assesses decussation (Figure 7.4). To maximize SNRs, the monitoring derivation’s input 1 should be the P37 maximum and input 2 the N37, usually CPc with normal decussation, CPi with nondecussation, and rarely Pz (8). Occasionally there is no apparent N37, and then an FPz input 2 may produce highest signal amplitude. However, by overwhelming this advantage, greater FPz noise often degrades SNR and slows feedback, so that FPz is infrequently used when the monitoring derivation is based on SNR (8). Note that Fz should not be used because it frequently contains anterior spread of the P37 field, causing in-phase signal cancellation, and it is likely also plagued by frontal EEG noise (10). One must confront the marked topographic variability of lower limb cortical SEPs to provide consistently optimal monitoring. This presents the complex problem of identifying optimal derivations. A proven technique for the left tibial SEP follows (8,9): 1. A referential recording of FPz, Cz, CPz, Pz, CP4, CP2, CP1, and CP3-mastoid along with bilateral popliteal fossa

recordings to confirm left stimulation is made in one panel and replicated. Simultaneously, a bipolar recording of Cz-CP4, CPz-CP4, Pz-CP4, CP1-CP4, CP3-CP4, Cz-FPz, CPz-FPz, Pz-FPz, CP1FPz, CP3-FPz, and Cz-Pz is made and replicated in another panel. 2. The P37 and N37 are evaluated in the referential recording during acquisition. A predominantly ipsilateral scalp P37 field confirms decussation; there is usually also a confirmatory contralateral N37. Reversed lateralization identifies nondecussation. 3. If nondecussation is disclosed, the recording is repeated using Cz-CP3, CPz-CP3, Pz-CP3, CP2-CP3, CP4-CP3, Cz-FPz, CPz-FPz, Pz-FPz, CP2-FPz, CP4-FPz, and Cz-Pz bipolar traces. If nondecussation is known preoperatively, this montage is used in step 1. 4. The bipolar trace showing fastest reproducibility (usually but not always also having highest signal amplitude) is selected for monitoring. The optimal input 1 (P37 maximum) and input 2 (N37) are usually but not always evident in the referential recording. In practice, both sides are simultaneously assessed using asynchronous parallel averaging and mirror-image bipolar montages (Figure 7.5). The procedure takes 5 to 10 minutes and is performed immediately after

FIGURE 7.4 Tibial SEP decussation assessment. Mastoid reference. A. Normal decussation. B. Nondecussation. CPz-FPz alone will miss nondecussation.

Husain 07

1/17/08

11:58 AM

Page 103

CH A P T ER 7: Ve r te bral Column Surge r y

•

1 03

FIGURE 7.5 Tibial P37 optimization. M, mastoid. Left and right results from different patients are shown for illustrative purposes. The ipsilateral P37 field and contralateral N37 confirm decussation. Cz-CP4 was optimal on the left and Pz-CP3 on the right. Both had larger signal amplitude, less EEG noise and substantially greater signal-to-noise ratio than CPz-FPz.

induction using a multichannel averager with templates for normal decussation and nondecussation. It was developed with a simpler eight-channel averager using sequential tibial nerve stimulation to obtain a bilateral referential recording and then selective bipolar comparative recordings based on the referential results. Table 7.1 lists the frequency of resulting optimal derivations. Note that commonly recommended routine monitoring derivations CPz-FPz, iCPi-FPz, and CPi-CPc are actually infrequently or rarely optimal. This approach optimizes lower limb cortical SEP recording for each side of each patient. In addition, one can confidently attribute P37 absence to pathology without being concerned about spurious absence due to topographic variation. Most importantly, it produces a mean 2.1:1 SNR advantage over the traditional CPz-FPz derivation (8). The arithmetic of averaging translates this to a

mean 4:1 acceleration of surgical feedback: the median sweep number to reproducibility is 128 for the optimized P37 and 512 for CPzFPz (8). Furthermore, for any given sweep number, mean amplitude variation due to residual noise is substantially reduced, so that true amplitude change is more readily detected and the chance of misleading random amplitude fluctuation is diminished (8). Although more complex than existing guidelines, the principles of patient-centered care justify the additional effort for superior results. If one chooses to apply only traditional FPz, CP4, CPz, and CP3 electrodes, some optimization can still be performed. Simultaneously recording CP4-, CPz-, and CP3-FPz will assess decussation. One could then compare CPz-CPc, CPi-CPc, CPz-FPz, and CPi-FPz for normal decussation or CPzCPi, CPc-CPi, CPz-FPz, and CPi-FPz for nondecussation and select the best of these for monitoring. Of these derivations, CPz-CPc

Husain 07

104

1/17/08

•

11:58 AM

Page 104

SE C T I ON I I : C l i n i c a l Me thod s

TABLE 7.1 Optimal Tibial P37 Monitoring Derivations for 314 Tibial Nerves of 157 patients* Input 2 (– down) Input 1 (+ down)

CPc

FPz

CPi†

Pz

Cz

58

3

1

3

CPz

119

32

5

0

Pz

40

12

0

–

iCPi

35

1

–

0

CPi

5

0

–

0

*CPc and CPi, CP4 or CP3 contralateral and ipsilateral to the stimulated nerve; iCPi, CP2 or CP1 intermediate centroparietal sites, ipsilateral. †Nondecussation. Source: Combined data from MacDonald et al. (8,9).

will most often but not always be best because it combines the most common P37 and N37 maxima and has lower noise than FPz derivations. If one insists on routinely applying a single standard derivation, then CPz-CPc is the best overall choice, but decussation should still be evaluated because CPz-CPi would be needed for nondecussation. In summary, the process for selecting the lower limb cortical SEP monitoring derivation is to (a) assess decussation, (b) compare derivations appropriate for the patient’s decussation status, and (c) use the derivation showing fastest reproducibility (highest SNR). The full optimization described above accounts best for known topographic variability.

Subcortical Potentials Subcortical SEP are noninvasive and show no significant topographic variability. Being generated in the cervicomedullary junction, they are applicable to spinal cord monitoring. Since this function is already served by cortical SEPs, the principal rationale for including subcortical SEPs is their lesser sensitivity to inhalation anesthesia. This was important when balanced inhalation anesthesia and suboptimal cortical SEP derivations were the rule, but it is not relevant to modern

cortical SEP methodology, which should include appropriate anesthesia and derivation selection. In fact, in this context, the low SNR of subcortical potentials interferes with feedback rapidity, particularly for lower limb monitoring (8). Consequently these potentials are presently not universally applied to spine surgery NIOM.

Upper Limb The upper limb subcortical P14-N18 complex is recorded with a scalp-noncephalic reference derivation. To avoid cortical N20/P22 contamination, the best input 1 is CPi for normal decussation or CPc for nondecussation. To avoid cervical cord potential contamination, EPc (Erb’s point contralateral to the stimulated nerve) is commonly used for input 2. Some use the mastoids instead. Others prefer a C5S-FPz derivation that combines the cervical N13 and inverted P14 to boost signal amplitude. All these derivations are prone to high noise content from long interelectrode distance, electrocardiogram (ECG), frontal-dominant EEG fast activity, and sometimes electromyography (EMG). This tends to degrade SNRs and prolong averaging. Nevertheless, upper limb subcortical potentials are relatively quickly reproduced, but they can be considered optional when cortical SEPs are properly recorded and present.

Lower Limb The tibial subcortical P31-N35 complex is recorded with a standard FPz-C5S or C2S derivation. Reversing the inputs produces an upgoing deflection that is sometimes incorrectly called cervical when FPz is actually the active electrode. Some prefer FPz-mastoid, which produces a slightly lower signal amplitude due to in-phase cancellation from some mastoid P31 activity. Owing to small signal (mean 1.1 µV) and high noise (mean 37 µV), the SNR of the P31 recorded with FPz-C5S is very low (mean 4.9 times lower than the optimized P37) (8). Thus, a median of 1,000 sweeps must be averaged for reproducibility,

Husain 07

1/17/08

11:58 AM

Page 105

CH A P T ER 7: Ve r te bral Column Surge r y

2,000 sweeps are needed for about 24% of tibial nerves, and replication is not practically demonstrable for about 7% (8). Clearly this potential is too slow and inconsistent for reliably effective monitoring. It might be a useful fallback potential when the P37 is absent due to cerebral pathology and when it is depressed by inhalation anesthesia or suboptimally recorded, but it should not be standard (8).

Peripheral Potentials Although not universally applied, peripheral SEPs enhance interpretation. The standard PF potential is satisfactory for lower limb monitoring. The traditional EP potential suffers from high noise content (particularly ECG). Instead, a brachial (Br) potential analogous to the PF and recorded medial to the biceps tendon just above the elbow is a better choice (4). Properly recorded peripheral SEPs require minimal averaging due to high SNRs that frequently make them visible in single sweeps (8). They provide a logical basis for determining supramaximal stimulus intensity. Moreover, they provide technical control and detect distal limb ischemia or pressure. When faced with cortical SEP decrement, simultaneous peripheral SEP loss immediately points to stimulus failure or distal limb problems, while preservation immediately excludes these factors.

Spinal Potentials Spinal SEPs are ascending dorsal column volleys recorded with invasive epidural, subdural, or interspinous ligament electrodes. They are almost immune to anesthesia because they are purely conductive (5,6). The latter property may be valuable when cortical SEPs are pathologically absent, depressed by inhalation anesthesia, or suboptimally recorded – particularly since spinal SEPs are more rapidly reproduced than the P31. Despite a large experience with spinal SEPs in Japan and England, the technique is not widely used (11,12). This may be due to an avoidance of invasive electrodes to eliminate the small chance of hemorrhage or infection

•

1 05

as well as to the evolution of modern noninvasive SEP methods.

Motor Evoked Potentials No standardized MEP monitoring approach or guideline currently exists. Nevertheless, MEPs should be a standard component of spinal cord monitoring during vertebral column surgery, and future guidelines will certainly include them. The basic technique is discussed in Chapter 2, and MEP monitoring has recently been extensively reviewed (13). Transcranial electrical stimulation (TES) activates the corticospinal system, and recordings are made from the spinal cord (D wave) and/or peripheral muscle.

Spinal D Waves Following single-pulse TES, a descending corticospinal tract volley known as the D (direct) wave can be monitored in the spinal epidural or subdural space. The bipolar recording electrode is inserted into the epidural space through a flavotomy after posterior spine exposure or threaded up the subdural space through lumbar puncture prior to opening. The D wave has several advantages. It is relatively immune to anesthesia because no synapses are involved and it does not require omission of neuromuscular blockade. In addition, it has rapid reproducibility due to high SNR and high stability. Furthermore, there is excellent correlation to long-term motor outcome for intramedullary spinal cord tumor and cerebral tumor surgery. On the other hand, there are several disadvantages in using the D wave for vertebral column surgery. The response is not clearly lateralized; excludes alpha motor neurons, which are more rapidly disabled by ischemia than tracts; and becomes too small to record at the lumbosacral cord where the corticospinal tracts end. In addition, the epidural technique cannot be used for anterior procedures or during opening and closure of poste-

Husain 07

106

1/17/08

•

11:58 AM

Page 106

SE C T I ON I I : C l i n i c a l Me thod s

rior spine surgeries. Furthermore, there is always the concern about invasive electrode complications, although rare. Most importantly, epidural D waves can produce false results during scoliosis surgery: an up to 75% decrease – or increase – of D-wave amplitude has been found in a number of patients despite unchanged muscle MEPs and neurologic outcome (14). This might be due to an increase or decrease of the distance between the epidural electrode and the spinal cord, as its position within the spinal canal shifts after straightening. Subdural D waves possibly might not pose this problem. Consequently, despite a large spine surgery D-wave monitoring experience in Japan and Australia, it seems unlikely that this technique will become a standard component of orthopedic NIOM (11,15). Note that it should be standard for intramedullary spinal cord tumor surgery (13).

Muscle Potentials Pulse train TES that evokes muscle MEP through temporal summation at alpha motor neurons is a major advance, having important advantages for spine surgery. No averaging is required owing to very high SNRs, so that feedback can be instantaneous (although interrupted by SEPs). It incorporates alpha motor neurons and assesses the corticospinal system from brain to individual limbs or even specific muscles relevant to the level of surgery. There is excellent correlation with early postoperative motor outcome, and muscle MEPs are more sensitive for spinal cord compromise than SEPs during spine surgery. The technique makes unpredicted spinal cord motor injuries unlikely and will almost certainly further reduce the incidence of spinal cord complications. Repetitive stimulation of the brain raises safety concerns; these are reviewed elsewhere (13,16,17). In short, the technique is sufficiently safe in expert hands using appropriate precautions. Bite injuries of the tongue or lips due to jaw muscle contractions are the most

common adverse effect, having an estimated incidence of 0.3%; there has been one instance of jaw fracture and one of endotracheal tube rupture (13). Preventive soft bite blocks are mandatory. Seizures, cardiac arrhythmias, and scalp burns are very rare. Muscle MEPs are even more sensitive to anesthesia than cortical SEPs. The main interference occurs at alpha motor neurons; anesthesia can obliterate muscle responses while having little effect on corticospinal volleys. Again, intravenous anesthesia such as propofol/opioid infusion appears optimal, but significant infusion increments or boluses can still reduce or obliterate muscle MEPs and may require increments of TES intensity and/or pulse number (13). Neuromuscular blockade is omitted after intubation or incomplete and tightly controlled. Stimulating electrodes are commonly placed at 10 to 20 central scalp sites approximately overlying motor cortex. Sites 1 or 2 cm anterior may diminish TES artifact by increasing the distance from the scalp SEP recording electrodes through which TES current reaches the headbox. The authors use C+1-cm sites and name them M (motor) sites. For the same reason, CP scalp SEP recording sites that are slightly posterior to C’ sites are preferred. Published stimulus parameters use three to eight rectangular pulses and a 1- to 5-ms interpulse interval (IPI). There is evidence that an IPI of 4 ms may generally be optimal for leg muscle MEPs, and the authors find five pulses with 4 ms IPI to be a good starting point for spine surgery (18). Various stimulators are effective, including specialized devices from Digitimer (www.digitimer.com), Inomed (www.inomed.com), Axon Systems (www.axonsystems.com), and Cadwell (www.cadwell.com) as well as standard Nicolet Endeavor or Viking NIOM stimulators (www.viasyshealthcare.com ) in constantvoltage mode (13). Stimulation is adjusted to clear suprathreshold responses of all targeted muscles or to the threshold of the last recruited muscle. Adjustments of pulse num-

Husain 07

1/17/08

11:58 AM

Page 107

CH A P T ER 7: Ve r te bral Column Surge r y

FIGURE 7.6 Scalp electrode sets. Solid circles designate the transcranial electric stimulation array and broken circles designate the scalp SEP recording set. [Modified from MacDonald (13), by permission of Springer.]

ber or IPI may be needed. Immediately preceding the test stimulus by one or more preconditioning stimuli facilitates the response and is a commonly used technique. It is helpful to apply an array of TES electrodes such as M4, M2, Mz, M1, and M3 from which stimulus montages can be selected to optimize technique (Figure 7.6). Because TES preferentially activates brain underneath

•

1 07

the anode, montages are given as anodecathode. Vertex-frontal TES can evoke bilateral leg MEPs, but is not very efficient and is ineffective for arm MEPs. Coronal TES that is more efficient and evokes arm and leg MEPs is more commonly used. Inter-hemispheric M3/M4 TES is very efficient, but tends to produce a strong patient twitch and may increase the likelihood of bite injury (13). M1/M2 TES can be an effective alternative with fewer tendencies to these problems, but some patients require M3/M4 for monitoring. Because of asymmetric muscle responses maximal contralateral to the anode, M1-M2 or M3-M4 TES with right MEP recording is alternated with M2-M1 or M4-M3 TES and left MEP recording. This strategy is reversed for nondecussation because responses are then maximal ipsilateral to the anode instead (3). Because hemispheric M3-Mz and M4-Mz TES produces predominantly unilateral corticospinal activation, these montages are used for decussation assessment (Figure 7.7). The authors do not recommend them for monitoring spine surgery because they are less efficient for leg MEPs (13). Thus, a way to optimize TES montages is to (a) assess decussation with M3-Mz and M4-Mz, (b) Compare M1/M2 and M3/M4 montages appropriate for the patient’s decussation status, and (c) favor M1/M2 montages if effective or use M3/M4 montages if needed (Figure 7.8).

FIGURE 7.7 Muscle MEP decussation assessment. APB = abductor pollicis brevis; AH = abductor hallucis. A. Normal decussation. B. Nondecussation.

Husain 07

108

1/17/08

•

11:58 AM

Page 108

SE C T I ON I I : C l i n i c a l Me thod s

FIGURE 7.8 Transcranial electric stimulation montage selection. Th = thenar; TA = tibialis anterior; AH = abductor hallucis. Nicolet Endeavor stimulator, 5 pulse trains, 4 ms IPI, 300 V. Hemispheric M3-Mz and M4-Mz stimuli confirmed motor decussation by showing anode-contralateral muscle MEP. Interhemispheric M3/M4 stimulation produced largest bilateral responses, maximal contralateral to the anode. M1/M2 stimuli produced sufficiently large leg MEP with less bilateral activation. M1-M2 and M2M1 were selected for monitoring right and left MEP. Some patients require M3/M4. [From MacDonald (13), by permission of Springer.]

INTERPRETATION AND WARNING CRITERIA Interpretation means to “set forth the meaning of.” Reliable NIOM interpretation requires appropriate training and experience in order to properly synthesize the relevant neurophysiologic, systemic, technical, and surgical factors.

Fundamental Considerations The wholly empiric traditional EP warning criteria of a more than 50% decrease of amplitude or 10% increase in latency need qualification. Latency prolongation is an illogical criterion because it is the hallmark of demyelination, a chronic process that is not expected during spine surgery. Rather, acute axonal conduction block or neuronal failure underlies intraoperative pathologic SEP or MEP decrement and predominantly reduce potential amplitude. Thus, amplitude is the primary intraoperative consideration. It is important to keep in mind that each successive SEP provides only an estimate of true potential amplitude distorted by residual noise that causes random trial-to-trial variation of this estimate. Poor reproducibility (greater than 30% amplitude variation,

unconvincing trace superimposition) compromises the identification of true amplitude change because large deviations of estimated amplitude can be due to chance alone; then it is best to rely on potential disappearance. Good reproducibility (20% to 30% variation, convincing superimposition) is essential to reliably identify true greater than 50% SEP amplitude decrements. With excellent reproducibility (less than 20% variation, nearly exact superimposition), lesser degrees of true SEP amplitude decrement can be detected and may be significant (4). The issue is whether there is an amplitude decrement that clearly exceeds random variation. Note that pathologic SEP decrements tend to underestimate the degree of sensory pathway conduction block due to central amplification. Muscle MEPs have no significant noise distortion but show substantial trial-to-trial variation due to inherent fluctuations of alpha motor neuron excitability (13). They also exhibit high sensitivity, so that pathologic decrements tend to overestimate the degree of corticospinal system compromise. Because of these properties, only disappearance or possibly marked attenuation amounting to virtual disappearance of a previously consistent muscle MEP is generally accepted as significant (13). This is not an excessively late warning

Husain 07

1/17/08

11:58 AM

Page 109

CH A P T ER 7: Ve r te bral Column Surge r y

•

1 09

during spine surgery because reappearance commonly follows intervention. Although there is supportive evidence for more sensitive criteria –such as a threshold elevation greater than 100 V MEPs or transformation from long-duration polyphasic to short-duration biphasic potentials—these approaches are presently controversial.(13)

Systemic Factors Systemic factors cause cortical SEP and/or muscle MEP alterations in virtually every surgery. Mild hypothermia prolongs latency but has less effect on amplitude; intentional deep hypothermia obliterates both potentials but is not used during spine surgery. Anesthetic increments can reduce amplitude, but even with stable anesthesia there is often unexplained progressive “potential fade” manifest by falling cortical SEP amplitudes and muscle MEP fading and/or increasing stimulus requirements (13,19). Modest blood pressure fluctuation does not normally cause EP changes because autoregulation maintains constant blood flow within the spinal cord and brain, but blood pressure and potential amplitude may rise and fall together with anesthesia that alters both. Systemic effects vary from minor to greater than 50% cortical SEP amplitude reduction, which occurs in at least 10% of spine surgeries, as well as striking muscle MEP reduction, threshold elevation, or even loss (4) (Figure 7.9). This undermines the traditional comparison to “baseline,” which incorrectly assumes a stable systemic state. Consequently, a different approach is needed for these systemically sensitive potentials in order to avoid false alarms. Systemic effects are generalized and tend to be gradual. Thus they produce approximately parallel four-limb EP changes that are reliably identified by upper limb controls during thoracic procedures (4). During cervical surgery, when four-limb EP decrement could be due to spinal cord compromise, gradual changes evolving over many minutes or hours

FIGURE 7.9 Unusually marked cortical somatosensory EP fade during scoliosis surgery. Plotted amplitudes are normalized to the largest observed values (1.0). Note the gradual generalized and approximately parallel amplitude reductions, making valid baseline determinations impossible. There was no injury.

are likely to be systemic. Large anesthetic boluses that can produce relatively abrupt generalized EP decrements should be avoided. Similarly, administration or potentiation of neuromuscular blockade can cause generalized muscle MEP loss (13).

Technical Factors Technical problems must also be identified. Peripheral SEP loss immediately points to the possibility of stimulus failure, which can then be rectified without surgical alarm. Similarly, simultaneous arm and leg MEP loss during thoracic procedures and loss of cranial muscle contractions or MEPs during cervical surgery suggests TES failure. Elevated TES or scalp SEP electrode impedance occasionally causes spurious potential reduction, identified by rechecking impedance. Amplifier saturation (flat line) of variable duration following electrosurgery or even TES can cause spurious EP reduction or loss. Resuming monitoring only after raw traces such as EEG from scalp

Husain 07

110

1/17/08

•

11:58 AM

Page 110

SE C T I ON I I : C l i n i c a l Me thod s

SEP derivations show amplifier recovery prevents this problem.

Pathologic EP Decrements After excluding systemic and technical factors, an abrupt focal EP decrement is the hallmark of pathologic compromise (4,13). These events typically occur within seconds or minutes, depending on the recording time resolution. In properly designed recordings, they are visually obvious, so that their interpretation does not really benefit from quantification.

Distal Nerve Disturbances Distal limb ischemia or pressure can cause EP decrement of the affected limb due to peripheral nerve conduction block and is readily identified by peripheral SEP loss. Resolution of these conditions might help to prevent peripheral nerve injury.

Proximal Nerve or Plexus Disturbances Proximal nerve or plexus conduction block occasionally causes EP decrement of the affected limb without affecting peripheral SEPs. This occurs frequently enough in the arms during spine surgery due to extreme shoulder positions that routine upper limb monitoring may also help prevent postoperative neuropathy or brachial plexopathy (4). Proximal nerve and plexus lesions have not yet been described in the legs.

Radiculopathy Due to radicular overlap, these NIOM methods do not reliably identify isolated root conduction block or injury (13). Dermatomal SEP monitoring might identify individual sensory root dysfunction but is controversial and must increase complexity and slow feedback. A discrete motor root injury may not be apparent in muscle MEPs, but sometimes does produce a visually obvious abrupt step reduction of MEP amplitude when the injured root supplies part of the recorded muscle’s innervation (13). It is therefore reasonable to monitor MEPs from several arm muscles spanning C4

through T1 roots during cervical surgery or from multiple leg muscles during lumbosacral surgery. There is evidence that concomitant EMG and pedicle screw stimulation may enhance lumbar nerve root protection.

Spinal Cord Compromise Thoracic spinal cord compromise produces abrupt bilateral or unilateral leg MEP loss and/or tibial cortical SEP amplitude decrement; an initially unilateral decrement can become bilateral but may remain asymmetric (4,13). Frequently there is no indication of impending MEP loss and the potential simply disappears in the next trial. Occasionally there is a retrospectively appreciable amplitude reduction immediately before, but muscle MEP instability makes it difficult to ascertain its significance until disappearance occurs. Four patterns of EP decrement are seen: (a) MEP only, (b) MEP followed by SEP, (c) simultaneous MEP/SEP, and (d) SEP only (rare). So far SEP followed by MEP decrement has not been reported but might be possible. These patterns confirm higher MEP than SEP sensitivity for cord compromise. Cervical cord compromise produces the same patterns but affecting both arm and leg potentials, depending on the level of involvement. Focal EP decrements can be seen to spread and such patterns further indicate pathologic change because systemic effects cannot explain them (Figure 7.10). Rapid surgical feedback assists interpretation by implicating the current or last surgical maneuver as the most likely cause of cord compromise. Undoing of this maneuver (e.g., releasing instrumentation, removing the last inserted hook, etc.) is frequently followed by EP restoration and no postoperative deficit, suggesting cord injury prevention. Because hypotension (mean arterial pressure less than 60 mmHg) can cause ischemia of a compromised spinal cord already having marginal perfusion, restoring blood pressure may be followed by potential restoration. Even without hypotension, raising mean arterial blood

Husain 07

1/17/08

11:58 AM

Page 111

CH A P T ER 7: Ve r te bral Column Surge r y

•

1 11

FIGURE 7.10 Evolving focal EP decrements. TA = tibialis anterior; AH = abductor hallucis. Traces are selected from the intraoperative record that had faster time resolution. The patient had nondecussation requiring reversed montages for monitoring. There was isolated left leg MEP loss at 16:08 immediately after T1 hook placement. Raising blood pressure had no beneficial effect. At 16:25 there was generalized MEP loss, and by 16:37 bilateral tibial SEP decrement. The hook was finally removed, followed by potential reappearance. Left leg muscle MEPs did not reappear until closure, and there was mild transient postoperative left leg weakness. The pattern suggests asymmetric high thoracic evolving to bilateral low cervical cord compromise and cannot be explained by systemic factors. [From MacDonald (13), by permission of Springer.]

pressure above 80 mmHg can sometimes appear to be effective and is a reasonable additional intervention. High-dose steroid administration may be considered but is not proven to improve outcome. Occurrences suggesting injury prevention should not be classified as false positive. Test condition analysis is inappropriate because intervention based on the test’s result influences the condition (outcome). This analysis can only be valid for the final EP results before patient awakening. Pathologic EP decrements persisting to the end of monitoring regularly predict clinical deficits that are congruent to the affected modality and limbs but not necessarily complete or permanent (13).

TECHNICAL CONSIDERATIONS Preparation Preparation begins with the monitoring request. The diagnosis and proposed surgical procedure indicate the structures at risk. The

patient’s neurologic status should be incorporated into monitoring expectations. Ideally, each patient would have a preoperative consultation including neurologic examination and EP studies, but this is not essential for neurologically intact patients. When resources permit, it is worthwhile doing this at least for those patients known to have neurologic impairment. Already absent SEPs are highly unlikely to be monitorable and the effort of attempting to do so can be avoided. However, some patients with impaired muscle strength and absent muscle MEPs to transcranial magnetic stimulation have monitorable TES MEPs. Patients with complete paraplegia or quadriplegia are an exception and should not undergo attempted MEP monitoring. Scoliosis patients with horizontal gaze palsy should have preoperative SEP and transcranial magnetic MEP decussation assessment whenever possible. Relative contraindications for TES must be noted, including epilepsy, cortical lesions, defects in the skull , raised intracranial pressure, cardiac disease, intracranial electrodes,

Husain 07

112

1/17/08

•

11:58 AM

Page 112

SE C T I ON I I : C l i n i c a l Me thod s

vascular clips, shunts, cardiac pacemakers, or other implanted biomedical devices (16). Because these are not proven to increase TES hazards, they must be carefully weighed against the risk of omitting MEPs, and a number of patients with these conditions have undergone uneventful TES MEP monitoring (13,17). Patients should be informed about monitoring procedures and all material hoped-for benefits and possible adverse effects including bite injury and the rarer possibilities of intraoperative seizure, cardiac arrhythmia, movement-induced injury, skin burns at electrode sites, and potential complications of any invasive electrodes to be used. Informed consent should be obtained. Monitoring equipment should be assembled. The recording instrument should have NIOM software and at least 8 channels; 16 or more channels are preferable. An integrated or external stimulator suitable for TES must be available. The necessary electrodes and supplies should be gathered. EEG cup electrode scalp sets can be prebraided to save time during application. Packaged sterile needle electrodes for scalp recording cannot be prebraided, but commercially available paired needles are convenient for distal stimulation and recording.

Procedure Hands should be washed and gloved for patient preparation (17). The head must be accurately measured. Recording leads destined for the same recording derivation must be tightly braided or paired to reduce extraneous electrical artifact. Impedances must be checked and should at least be less than 5 kOhms; a 2-kOhm limit is preferable to ensure balanced impedance values. The application of surface electrodes before the patient’s call to surgery minimizes impact on operating room and anesthesia time. Proper skin preparation with light abrasion regularly produces low impedance. A few

patients unable to tolerate preparation while awake need postinduction application. Prebraided EEG cup scalp electrode sets for SEP recording and TES are effective and securely attached with collodion. Standard adhesive ECG discs are effective for shoulder ground and limb SEP stimulation and recording sites. Their snap-on reusable leads can be prebraided and labeled for rapid connection. Patients with unusually thick or edematous ankles may require needles for tibial nerve stimulation. Stimulating bar electrodes can cause pressure necrosis and are discouraged (17). Surface electrodes are suitable for muscle MEPs, but paired subdermal needles inserted after induction are widely preferred. The routine application of all electrodes in the operating room after induction is a common alternative. The quickness of needle electrodes minimizes preparation time, which becomes an issue in this setting. Spiral needles fix securely in the scalp; straight needles are more difficult to secure here but are readily fixed with tape elsewhere. Although surface electrodes are more timeconsuming and require skill in the use of collodion, they have excellent recording and safety profiles. Needle electrodes pose a greater risk of accidental electrosurgey-induced skin burns, infectious complications, needle-stick events, and broken tips lodged under the skin (17). They are generally discouraged for neurophysiologic testing but are convenient for NIOM. Both are effective and monitoring teams should decide which best matches their circumstances and patients’ needs. The anesthesiologist should be consulted for assistance in obtaining optimal monitoring. Ideally this will involve intravenous anesthesia if possible. Neuromuscular blockade may be necessary for intubation but should be short-acting to avoid interfering with initial muscle MEP recordings. A courteous reminder to place a soft bite block may be needed. The first traces after induction should identify optimal derivations and assess decussation if not already done. The authors begin

Husain 07

1/17/08

11:58 AM

Page 113

CH A P T ER 7: Ve r te bral Column Surge r y

with tibial cortical SEP optimization that simultaneously assesses sensory decussation. Then initial upper limb SEP can be obtained without needing to reassess decussation. The first MEP should be recorded bilaterally to M3-Mz and then M4-Mz TES to assess motor decussation if not already known. Then right and left MEP are obtained to M1/M2 or if necessary, M3/M4 stimuli according to the patient’s decussation status. Muscle MEPs may initially be absent if inhalation agents have been used for induction or there is residual neuromuscular blockade; one may need to wait until a favorable state for MEP recording to be established. Initial baselines are set after determining optimal technique. During cervical procedures, this should be accomplished prior to patient positioning. The anesthesiologist’s collaboration in quickly establishing a suitable systemic state for SEP/MEP recording is particularly important for these patients. One or

•

1 13

more EP sets should be obtained after positioning. Monitoring is suspended during the electrosurgery of opening and then resumed. Baselines are reset after opening, and then sequential EP sets (upper and lower limb SEPs, right and left MEPs) are acquired as rapidly as permitted by the averaging time needed for SEP reproducibility. Spinal cord ischemia causes muscle MEP loss within 2 minutes and is an important injury mechanism during spine surgery (4,13). Therefore set acquisition within about 2 minutes without compromising reproducibility should be the goal (4,8). SNR-based SEP optimization allows this in the majority of patients (Figure 7.11). Pathological EP decrements commonly arise within this time frame (Figure 7.12). The routine use of lower SNR derivations such as CPz-FPz and FPz-C5S will very often exceed 2 minutes acquisition time or impair accuracy if averaging is cut short of reproducibility (8).

FIGURE 7.11 A typical example of rapid surgical feedback made possible through SNR-based SEP optimization. The average feedback interval was 1.4 minutes. Br = median brachial peripheral SEP; Th = thenar; TA = tibialis anterior; AH = abductor hallucis.

Husain 07

114

1/17/08

•

11:58 AM

Page 114

SE C T I ON I I : C l i n i c a l Me thod s

FIGURE 7.12 Typically abrupt focal decrements. Th = thenar; TA = tibialis anterior; AH = abductor hallucis. The average feedback interval was 1.5 minutes. There was no sign of trouble at 15:10:30 while a high thoracic sublaminar hook was being inserted, but 1 minute and 18 seconds later there was right leg MEP loss and marked tibial SEP decrement. One minute and 9 seconds after that, decrements were bilateral. The hook was removed and potentials started reappearing in a few minutes. Left leg potentials showed full SEP and partial MEP recovery; there was no clinical deficit of this limb. Right leg potentials showed partial SEP recovery and persistent MEP absence; there was moderate postoperative leg weakness and dorsal column sensory deficit but preserved ambulation and eventual complete clinical recovery. This suggested probable spinal cord contusion. [From MacDonald (13), by permission of Springer.]

Baselines can be reset if systemic potential fading is observed. Maintaining the presence of fading muscle MEP may require intensity and/or pulse number increments, more facilitation with repetitive preconditioning pulse trains, and sometimes a switch from M1/M2 to M3/M4 (13). Patient twitching with TES normally does not interfere with spine surgery; surgeons quickly become accustomed to it and need not be warned before stimulation (4). Anterior

cervical surgery is an exception, because neck muscle contractions may interfere and careful stimulus timing and communication may be required (13). Documentation should be thorough. Anesthesia, blood pressure, temperature, TES parameter or other technical changes, surgical maneuvers, and communications with anesthesiology and surgery teams must be noted. The surgeon need not be informed about systemic changes unless they begin to prevent

Husain 07

1/17/08

11:58 AM

Page 115

CH A P T ER 7: Ve r te bral Column Surge r y

effective monitoring, in which case the anesthesiologist must also be consulted for help. Technical problems must be corrected without surgical alarm. The likely meaning of and possible interventions for pathologic decrements must be brought to surgical attention immediately; a competent intraoperative neurophysiologist should be responsible for this call and the surgeon’s response and interventions should be documented. Every effort must be made to avoid false warnings or reassurance. These can adversely affect the patient’s surgical result or neurologic outcome and seriously undermine the surgeon’s confidence (17). Such events should be rare with good technique.

Postprocedure Monitoring is normally discontinued at closure. Rarely, a patient is kept under sedation postoperatively; then the possibility of extended intensive care unit monitoring can be considered. Unless sedation is very deep, TES may be too painful to apply in this circumstance. Hands should be washed and gloved for disassembly (17). Collodion-fixed electrodes are removed with acetone. Neither of these flammable substances should be in open use during electrosurgery (17). EEG cup electrodes are bagged for later cleaning and disinfecting (17). Needle electrodes are handled by the stem, not recapped, and disposed of in a sharps container (17). Adhesive and invasive electrodes are disposed of in the appropriate biomedical hazard waste container. Any sign of skin burn or injury apart from minor abrasion from skin preparation should prompt an investigation to rectify the cause (17). Subdermal broken needle tips must be brought to the surgeon’s attention for subsequent surgical removal. The monitoring instrument and its cables and boxes are cleaned with disinfectant. Standard maintenance procedures must be followed, including inspection and leakage

•

1 15

current testing at least every 6 months, at any sign of malfunction, and after any repair (17). The patient’s outcome should be correlated with NIOM results. Any discrepancies should be thoroughly evaluated, and this may include postoperative EP studies. It is important that monitoring, surgical, and anesthesiology teams understand the relationship of their NIOM results to patient outcome.

CONCLUSIONS Vertebral column surgeries risk spinal cord, nerve root, brachial plexus, and peripheral nerve injury. SEP monitoring halves the risk of spinal cord injury, but the possibility of discrete corticospinal or dorsal column compromise makes the addition of muscle MEP monitoring important. Modern techniques— including intravenous anesthesia, SNR-based SEP derivation selection, and decussation assessment—optimize recording for each patient. These methods provide four-limb EP surgical feedback within 2-minute intervals for the majority of patients to enhance surgical correlation and the prevention of injury. Systemically sensitive cortical SEPs and muscle MEPs do not conform well to percentage of baseline amplitude criteria. Instead, gradual, generalized EP changes are systemic, while abrupt focal decrements clearly exceeding trial-to-trial variation identify pathologic change. The addition of TES muscle MEP monitoring will almost certainly further reduce the risk of spinal cord injury because of its greater sensitivity and motor specificity.

REFERENCES 1. Nuwer MR, Dawson EG, Carlson LG, et al. Somatosensory EP spinal cord monitoring reduces neurologic deficits after scoliosis surgery: results of a large multicenter survey. Electroencephalogr Clin Neurophysiol 1995; 96:6–11.

Husain 07

116

1/17/08

•

11:58 AM

Page 116

SE C T I ON I I : C l i n i c a l Me thod s

2. Bosley TM, Salih MA, Jen JC, et al. Neurologic features of horizontal gaze palsy and progressive scoliosis with mutations in ROBO3. Neurology 2005;64:1196–1203. 3. MacDonald DB, Streletz L, Al-Zayed Z, et al. Intraoperative neurophysiologic discovery of uncrossed sensory and motor pathways in a patient with horizontal gaze palsy and scoliosis. Clin Neurophysiol 2004;115:576–582. 4. MacDonald DB, Al-Zayed Z, Khodeir I, Stigsby B. Monitoring scoliosis surgery with combined transcranial electric motor and cortical somatosensory EP from the lower and upper extremities. Spine 2003;28:194–203. 5. Burke D, Nuwer MR, Daube J, et al. Intraoperative monitoring. The International Federation of Clinical Neurophysiology. Electroencephalogr Clin Neurophysiol Suppl 1999;52:133–148. 6. American Electroencephalographic Society. Guideline eleven: guidelines for intraoperative monitoring of sensory EP. J Clin Neurophysiol 1994;11:77–87. 7. Toleikis JR. American Society of Neurophysiological Monitoring. Intraoperative monitoring using somatosensory EP. A position statement by the American Society of Neurophysiological Monitoring. J Clin Monit Comput 2005;19:241–258. 8. MacDonald DB, Al Zayed Z, Stigsby B. Tibial somatosensory EP intraoperative monitoring: recommendations based on signal to noise ratio analysis of popliteal fossa, optimized P37, standard P37 and P31 potentials. Clin Neurophysiol 2005;116:1858–1869. 9. MacDonald DB, Stigsby B, Al-Zayed Z. A comparison between derivation optimization and Cz’-FPz for posterior tibial P37 somatosensory EP intraoperative monitoring. Clin Neurophysiol 2004;115:1925–1930. 10. MacDonald DB. Individually optimizing posterior tibial somatosensory EP P37 scalp deri-

11.

12.

13.

14.

15.

16.

17.

18.

19.

vations for intraoperative monitoring. J Clin Neurophysiol 2001;18:364–371. Iwasaki H, Tamaki T, Yoshida M, et al. Efficacy and limitations of current methods of intraoperative spinal cord monitoring. J Orthop Sci 2003;8:635–642. Forbes HJ, Allen PW, Waller CS, et al. Spinal cord monitoring in scoliosis surgery. Experience with 1168 cases. J Bone Joint Surg Br 1991;73:487–491. MacDonald DB. Intraoperative motor EP monitoring: overview and update. J Clin Monit. 2006; EPub ahead of print July 11, 2006. Ulkatan S, Neuwirth M, Bitan F, et al. Monitoring of scoliosis surgery with epidurally recorded motor EP (D wave) revealed false results. Clin Neurophysiol 2006;117: 2093–2101. Burke D, Hicks RG. Surgical monitoring of motor pathways. J Clin Neurophysiol 1998; 15:194–205. MacDonald DB. Safety of intraoperative transcranial electric stimulation motor EP monitoring. J Clin Neurophysiol 2002;19:416–429. MacDonald DB, Deletis V. Safety issues during surgical monitoring. In: Nuwer MR, ed, Intraoperative Monitoring of Neural Function: Handbook of Clinical Neurophysiology. 2008 (in press). Deletis V, Rodi Z, Amassian VE. Neurophysiological mechanisms underlying motor EP in anesthetized humans. Part 2. Relationship between epidurally and muscle recorded MEP in man. Clin Neurophysiol 2001;112:445–452. Lyon R, Feiner J, Lieberman JA. Progressive suppression of motor EP during general anesthesia: the phenomenon of “anesthetic fade.” J Neurosurg Anesthesiol 2005;17:13–19.

Husain 08

1/17/08

11:59 AM

Page 117

8

Spinal Cord Surgery

Thoru Yamada Marjorie Tucker Aatif M. Husain

T

potentials (MEPs) evaluating the anterior spinal cord and motor pathways have become available. There are several methods for SEP and MEP monitoring. Which method is chosen depends on the type of surgery and the preference of the neurophysiologist. Each method has advantages and disadvantages. The NIOM laboratory should be able to perform any alternative method if one method does not work. Since SEPs and MEPs are complementary and combining them extends the coverage of spinal cord functions, both types of monitoring should be performed whenever possible. Using both types of NIOM will also help in cases when one modality fails to yield measurable or consistent responses.

he neurophysiologic intraoperative monitoring (NIOM) of spinal cord function has been used for spinal fusion and instrumentation surgeries by orthopedic surgeons, surgeries of thoracic and thoracoabdominal aorta and cardiac surgeries requiring cardiopulmonary bypass by cardiothoracic surgeons, coarctation repair by vascular surgeons, and surgeries of intramedullary and extramedullary tumors and arteriovenous malformations (AVMs) of the spinal cord by neurosurgeons. This chapter focuses on NIOM in the latter procedures. Most surgeries on the spinal cord are for the resection of tumors. During surgery, spinal cord injury may occur either by direct physical damage or by ischemic insult to the spinal cord. In orthopedic procedures the risk of physical damage is more likely, and ischemia often occurs in cardiac and vascular surgery ischemia; however, both can occur in neurosurgical procedures. Historically, the standard for NIOM of spinal cord function has been the monitoring of somatosensory evoked potentials (SEPs). However, SEPs evaluate primarily posterior spinal cord function, and false-negative findings (i.e., anterior spinal cord damage and paraplegia occurring without being detected by SEPs) have been reported.(1,2) More recently, transcranial electrical motor evoked

BASIC ANATOMY OF THE SPINAL CORD Spinal Cord Anatomy The spinal cord is protected by multiple vertebrae, passing through the vertebral foramina of 7 cervical, 12 thoracic, 5 lumbar, and 5 sacral vertebrae. The spinal cord is the downward extension of the medulla oblongata and ends at the conus medullaris at the level of the T12 spine. From the conus, a bundle of nerve fiber called the cauda equina 117

Husain 08

118

1/17/08

•

11:59 AM

Page 118

SE C T I ON I I : C l i n i c a l Me thod s

extends down to the coccyx and ends as the filum terminale. The dura and arachnoid (subarachnoid space) mater extend down to the level of S2 vertebra. There are 31 pairs of spinal nerves: 8 cervical, 12 thoracic, 5 lumbar, 5 sacral, and 1 coccygeal. Each spinal nerve passes through an intervertebral foramen between two vertebrae. In the cervical spine, the C1 spinal nerve passes above the C1 vertebra and the C8 spinal nerve passes below the C7 vertebra (i.e., between the C7 and T1 vertebrae). In the thoracic, lumbar, and sacral spine, each spinal nerve is numbered in accordance with the number of the vertebra just above it. For example, the T1 spinal nerve passes below the T1 vertebra and the L5 lumbar spinal nerve passes below the L5 vertebra. The nerves of the upper cervical spinal lie horizontally; further caudally, however, the spinal nerves assume an increasingly oblique and downward direction as they proceed to their foramina to exit the spinal canal. The C7 spine level is identified by a protruded spinous process at the lower part of the neck when the neck is flexed. The T12 spine level is identified by the lowest rib, while the L4 spine level corresponds with the horizontal line between the left and right upper borders of iliac crest.

Fiber Tracts of the Spinal Cord The spinal cord consists of gray matter, which occupies the central portion of spinal cord, and white matter surrounding the gray matter. The gray matter is made up of collections of cell bodies, dendrites, axons, and axon terminals; the white matter consists of axons of longitudinally running nerve fiber tracts. Only the anatomic structures relevant to spinal cord NIOM are discussed here. The spinal cord comprises two major pathways. One is an afferent (sensory) pathway, which carries information from the peripheral receptors to the brain, and the other is an efferent (motor) pathway, which conveys messages from the brain to the periphery. The afferent

pathway starts from peripheral receptors, ascends in the sensory nerve fibers, and enters the spinal cord via the dorsal root. In the spinal cord, the sensory fibers carrying temperature and pain (superficial sensation) via unmyelinated or small myelinated fibers cross to the opposite side and ascend in the spinothalamic tract, which is located in the anterior and lateral aspect of the spinal cord. The sensory fibers carrying position and vibration sensation via large myelinated fibers enter the dorsal column, which occupies the posterior portion of the spinal cord. These fibers then ascend on the same side to the brainstem. The dorsal column consists of two fasciculi: the fasciculus gracilis carries fibers from the lower limb and the fasciculus cuneatus carries fibers from the upper limb. The fasciculus gracilis and fasciculus cuneatus reach the brainstem and synapse on the gracilis nucleus and cuneatus nucleus. The fiber tract then crosses to the opposite side, forming the medial lemniscus. There the spinothalamic tract carrying superficial sensation and the fasciculi gracilis and cuneatus carrying deep sensation join together and ascend through the brainstem contralateral to side that is stimulated. Eventually the fiber tracts reach the thalamus and synapse on the ventral posterolateral (VPL) nucleus. From here the sensory fibers pass through the posterior limb of internal capsule and reach the sensory cortex at the postcentral gyrus. The sensory cortex representing the upper limb is located on the lateral convexity of the hemisphere, and the lower limb is represented at the vertex and mesial aspect of the hemisphere. It is important to note that SEPs elicited by electrical stimulation are primarily mediated by the dorsal column pathway and posterior spinal cord, not the spinothalamic tract or anterior spinal cord. An abnormal SEP, therefore, is seen primarily in patients with impaired proprioception but not in those with impaired pain and temperature sensation. The main efferent pathway is the corticospinal tract, which starts from the motor cortex at the precentral gyrus. It descends via

Husain 08

1/17/08

11:59 AM

Page 119

CH A P T ER 8: Spinal Cord Surge r y

the posterior limb of the internal capsule to the pyramidal decussation of the brainstem, where a major part of tract crosses to the opposite side. The tract continues to descend in the lateral aspect of the spinal cord and finally reaches the anterior horn cells. After synapsing on the anterior horn cells, the motor fibers exit the spinal cord via ventral roots and end at the motor points of muscle fibers.

•

1 19

lateral aspect of the entire length of the spinal cord. They receive blood supply from the vertebral arteries with contributions from posterior radicular arteries. In the cervical region, the radicular arteries arise from deep cervical and subclavian arteries. At thoracic, lumbar, and sacral regions, posterior radicular arteries arise from posterior intercostal, lumbar, and lateral sacral arteries.

Arteries of the Spinal Cord The spinal cord is supplied by one anterior and two posterior spinal arteries. For details on vascular supply of the spinal cord, the reader is referred to the Chapter 15. An overview is presented here. The anterior spinal artery runs the entire length of spinal cord in the midline. It is supplied by the subclavian and vertebral arteries superiorly. In the cervical and upper thoracic region, it is supplied by one or two radicular arteries. In the lower thoracic to upper lumbar region, it receives blood supply from the arteria radicularis magna (ARM), also known as the artery of Adamkiewicz. This large artery is responsible for supplying much of the lower part of the spinal cord. Branches of the ARM also synapse with the posterior spinal arteries. The posterior spinal arteries are paired arteries that travel in parallel on the postero-

SPINAL CORD TUMORS Most surgeries on the spinal cord are for resection of tumors. Approximately 2% to 4% of all primary central nervous system tumors occur in the spinal cord. This section offers a brief overview of these tumors—their presentation, diagnosis, and surgical management.

Classification Spinal cord tumors are most often classified based on their location. Extradural tumors arise outside the dura mater, most often from the vertebral bodies. Intradural extramedullary tumors arise from within the dura mater but outside the spinal cord. Finally, intramedullary tumors arise from within the spinal cord. A classification of these tumors is presented in Table 8.1.

TABLE 8.1 Classification System for Spinal Cord Tumors Intradural, Extramedullary

Extradural Malignant

Intramedullary

Benign

Osteosarcoma

Osteoid osteoma

Meningioma

Astrocytoma

Chondrosarcoma

Osteoblastoma

Schwannoma

Ependymoma

Leiomyosarcoma

Osteochondroma

Neurofibroma

Oligodendroglioma

Ewing’s sarcoma

Chondroblastoma

Metastases

Ganglioglioma

Chordoma

Giant cell tumor

Lipoma

Multiple myeloma/

Hemangioma

solitary plasmacytoma Metastases

Aneurysmal bone cysts

Neuroblastoma Hemangioblastoma Vascular lesions Metastases

Husain 08

120

1/17/08

•

11:59 AM

Page 120

SE C T I ON I I : C l i n i c a l Me thod s

Extradural Tumors Extradural tumors can arise either from the bone or soft tissues surrounding the spinal cord, or they may be metastatic (3). Primary tumors of the bone and soft tissue are rare and can be malignant. Osteosarcomas arise from bony structures and may be seen in patients with Paget’s disease. Chondrosarcomas are tumors of cartilage and usually originate in the vertebral body. Leiomyosarcomas arise from mesenchymal elements such as muscle cells. Ewing’s sarcomas can arise in the bone or soft tissue and are small, round, blue-cell tumors. Chordomas arise from remnants of the notochord and occur most often near the clivus, cervical spine, and lumbosacral region. Multiple myeloma or a solitary plasmacytoma can occur in the spine; these are malignancies of plasma cells. Metastatic tumors can also involve the spine. Any malignant tumor can metastasize to the spine, but the commonest primaries are prostate, breast, and lung. A variety of benign lesions can also occur in the extradural space. Osteoid osteomas are small lesions (less than 2 cm in diameter) that typically arise from spongy bone and occur in the posterior aspect of vertebral bodies. Osteoblastomas are similar to osteoid osteomas but are larger (greater than 2 cm) and consequently more likely to produce neurologic symptoms. Osteochondromas consists of healthy bone with a cartilaginous cap and occur most often in the cervical spine. Chondroblastomas are tumors of immature cartilage and only rarely occur in the spine. Giant-cell tumors are very vascular and usually affect the vertebral body; they have a high recurrence rate after resection. Vertebral hemangiomas are nonneoplastic lesions consisting of thin-walled blood vessels and are often incidental findings. Aneurysmal bone cysts are nonneoplastic expansile bloodfilled cysts commonly located along the posterior aspect of the thoracolumbar spine; they cause symptoms by their local expansion.

Intradural Extramedullary Tumors Intradural extramedullary tumors make up the majority of intradural spinal cord

tumors (4). Meningiomas and nerve sheath tumors are the commonest types. Meningiomas arise from arachnoid cells and are adherent to the dura mater. They occur most often in the thoracic spine. Rarely, meningiomas can be extradural, but in these situations they almost always have an intradural extension. Schwannomas and neurofibromas are nerve sheath tumors. They typically arise from spinal nerve roots before they leave the dural sac. Occasionally they extend beyond the dural sheath, making them dumbbell-shaped. Schwannomas are usually solitary lesions, whereas neurofibromas are often multiple and seen in neurofibromatosis type I. Intradural, extramedullary metastatic tumors are rare and are often the result of drop lesions of systemic or intracranial neoplasms. They occur most often in the thoracic spine. Other types of intradural, extramedullary tumors are rare and typically located in the cauda equina; they are not discussed further.

Intramedullary Tumors Intramedullary tumors are seen more often in children than in adults. Astrocytomas are the commonest type of intramedullary tumor and are most often located in the cervicothoracic region. Pilocytic astrocytomas are low-grade, well-circumscribed neoplasms that are amenable to near complete resection. Fibrillary astrocytomas are more likely to be high-grade neoplasms; they are diffuse lesions. They cannot be resected in their entirety. Intramedullary ependymomas are more common in adults than children and can occur anywhere in the spinal cord. There is a discrete plane between the ependymoma and adjacent spinal cord, allowing for easier surgical dissection. Oligodendrogliomas are infiltrating tumors located mostly in the thoracic spine and are often associated with a syrinx. They frequently have leptomeningeal metastases and total resection is not possible. Gangliogliomas contain neurons and glial cells and are most often located in the cervi-

Husain 08

1/17/08

11:59 AM

Page 121

CH A P T ER 8: Spinal Cord Surge r y

cothoracic region. They can often be removed in total. Neuroblastomas contain neuroblasts, ganglion cells, and Schwann cells and are located mostly in the cervical spine. Lipomas consist of fatty tissue mixed with neuronal and connective tissue and are seen mostly in the cervicothoracic region. They are infiltrating tumors and treatment is aimed at debulking rather than complete removal. Hemangioblastomas are highly vascular tumors whose feeding arteries and draining veins can be seen on magnetic resonance imaging (MRI). They are most often present in the cervicothoracic region and are associated with von Hippel–Lindau syndrome. Complete resection of these tumors is possible. Vascular lesions of the spinal cord are considered AVMs and not true neoplasms. Those with small dominant vessels are called capillary hemangiomas, whereas those with larger vessels are called cavernous hemangiomas. They are most commonly seen in the cervicothoracic region, but multiple lesions are not infrequent. Complete surgical resection is possible, although endovascular treatment is also sometimes considered. Intramedullary metastases are uncommon; when they occur, the primary is most likely to be lung. These are rarely treated surgically.

Clinical Presentations Spinal cord tumors can be silent for many years. Onset is usually insidious, the most frequent complaint being back and neck pain. The pain is usually worse in the horizontal position and has a diffuse aching character rather than a radiating one. Many patients, especially those with a thoracic lesion, have a spinal deformity. Paresthesias are also common and are often asymmetric. Motor complaints occur after sensory symptoms. This may be manifest as clumsiness and falling in adults and as motor regression in children. Spasticity, hyperreflexia, and Babinski signs will also be present. Flexor spasms can also occur, causing

•

1 21

considerable pain. Sphincter dysfunction occurs late in the course of disease progression. The presentation is highly dependent on the location of the tumor. A tumor located in the posterior spinal cord is much more likely to causes dorsal column dysfunction prior to weakness, whereas the opposite is true for a more anteriorly located lesion. In most cases of spinal cord tumors, progression of symptoms is insidious owing to gradually worsening compression of neural structures. Sudden worsening often implies a vascular compromise and is often not reversible. When symptoms are slowly progressive, treatment outcome is likely to be much better than when the worsening is acute.

Diagnostic Evaluation The diagnostic study of choice for spinal cord tumors is MRI. This must be obtained with and without a contrast agent (gadolinium). The T1 sequences are useful in showing solid parts of the tumor, whereas T2 sequences demonstrate the cerebrospinal fluid (CSF) and cystic components of the tumor. Although MRI cannot make histologic diagnoses, the characteristic appearance of lesions allow a reasonably accurate classification. Myelography and computed tomography (CT) myelography were the “gold standard” in imaging spinal cord lesions prior to MRI. These tests are now reserved for exceptional situations in which an MRI cannot be obtained. Lesions of the bone may be better visualized by CT scans. Plain x- rays can be helpful if the tumor causes lytic bone lesions. Angiography may be helpful with hemangioblastomas and hemangiomas.

Surgery Surgery for spinal cord tumors is preformed with the patient in the prone position, although some surgeons prefer a flexed crouch position. The incision is usually lim-

Husain 08

122

1/17/08

•

11:59 AM

Page 122

SE C T I ON I I : C l i n i c a l Me thod s

ited to two vertebral levels over the focal point of the tumor. Dissection is carried to the lamina and then either a laminectomy or hemilaminectomy is performed, depending on the location of the tumor. Extradural tumors can be removed or debulked without opening the dura mater. Intradural tumors can often be seen even prior to opening the dura mater. A longitudinal incision is used to open the dura mater over the site of the tumor. For extramedullary tumors, the arachnoid mater over the tumor is opened and the tumor is entered. Intracapsular debulking of the tumor is achieved with blunt dissection or with alternative techniques such as the cavitronic ultrasound surgical aspirator (CUSA) or the carbon dioxide (CO2) laser. Once the tumor is debulked, the capsule collapses and can be removed from the dura mater. Intramedullary tumors can be localized in the surgical field with ultrasound. These tumors are usually approached through a midline incision of the spinal cord between the dorsal columns (myelotomy). For more laterally placed lesions, the approach can be through the dorsal root entry zone. The myelotomy may cause significant changes in the NIOM, as discussed below. Dissection of the tumor depends on its suspected pathology. Astrocytomas are dissected from inside out until the cord–tumor interface is reached. Ependymomas are dissected from one pole to the other, as a cleavage plane can be found in these tumors. Lipomas are well demarcated but tightly adherent to the spinal cord; consequently complete resection is not possible. Dissection of these tumors can be aided with a CUSA or CO2 laser. The spinal cord is then reapproximated with sutures. Once the tumor is removed and hemostasis is achieved, if the dura mater was opened, it is sutured in a watertight fashion. The thecal sac is reexpanded with fluid to prevent hemorrhage. The paraspinal muscles, fascia, and skin are then approximated and closed in layers.

MODALITIES AND METHODS OF NIOM Somatic Evoked Potentials There are several methods for SEP monitoring. The most commonly used method is scalp SEP recording after stimulation of peripheral nerves—for example, ulnar or median nerve stimulation for the upper limb and tibial or peroneal nerve stimulation for the lower limb. This is the gold standard for SEP monitoring and is been discussed elsewhere in this chapter. Here alternative methods are be described.

Peripheral Nerve Stimulation with Recording from the Spinal Cord After stimulation of a peripheral nerve in the lower extremity, a spinal potential can be recorded from a pair of electrodes embedded in a soft, flexible wire placed in the epidural space of the high thoracic or cervical spinal cord. The waveform becomes smaller and more polyphasic the more rostral the level of recording (5) (Figure 8.1). The onset latency of the waveform is dependent on the location of the recording electrodes. In general, the onset latency in adults at the low cervical or high thoracic spine is about 25 to 30 ms with stimulation in the lower extremity. There are several advantages to this recording method. Because the response is robust, less averaging is required than scalp-recorded SEPs. This response is affected little by anesthetics, since it is oligosynaptic. Finally, the waveform can be recorded at multiple levels, which allows determination of the segmental level of damage if multiple electrodes are used.

Spinal Cord Stimulation and Spinal Cord Recording Two sets of epidural electrodes are placed both rostral and caudal to the surgical site. Either pair of electrodes can be the recording or stimulating site. A single stimulus usually yields a robust response consisting of two negative peaks (NI, NII) with a latency of less than 7 ms (5) (Figure 8.2).

Husain 08

1/17/08

11:59 AM

Page 123

CH A P T ER 8: Spinal Cord Surge r y

•

1 23

FIGURE 8.1 Spinal evoked potential (SpEp) after stimulation of a peripheral nerve (left) and after stimulation of the spinal cord (right). After stimulation of tibial nerve, SpEps are recorded at the T10 and T3 spinal levels. Note greater polyspike waves with the higher level of recording. After stimulation of the spinal cord at T3 spine, SpEps recorded at the T10 spine showed W-shaped waveform consisting of NI and NII potentials. Reversing the stimulating and recording electrodes will record a similar waveform. [Reproduced with permission from Machida et al. (5).]

Among the advantages of this method is that the stimulus intensity required is much lower, usually less than 5 mA, compared to peripheral nerve stimulation, where 20 to 40 mA is used. It is not dependent on peripheral nerve function and can be applied to the patient who has a peripheral neuropathy or whose peripheral nerves cannot be adequately stimulated. Additionally, in this technique no averaging is required, and it is not affected by anesthetics.

Spinal Cord Stimulation and Scalp Recording Instead of stimulating peripheral nerves, essentially the same waveform but with a

much shorter latency can be recorded from the scalp after stimulation of the caudal spinal cord via an epidural electrode (Figure 8.2). The stimulus intensity is less than 10 mA, and usually less than 50 responses need to be averaged, which is much less than required for peripheral nerve stimulation. The advantage of this method is that it is not dependent on peripheral nerve conduction and can be used for patients with a peripheral neuropathy or those whose peripheral nerves are not accessible. It has a shorter acquisition time compared to the usual method of obtaining SEPs. Additionally it can be used for monitoring cervical cord function, while spinal cord stimulation and recording are primarily

Husain 08

124

1/17/08

•

11:59 AM

Page 124

SE C T I ON I I : C l i n i c a l Me thod s

FIGURE 8.2 Scalp-recorded SEPs and SpEps after stimulation of the spinal cord. With stimulation of the spinal cord at the T3 spine level (right), the scalp-recorded far-field potential of P31 showed an expectedly shorter latency than the response to stimulation at the T3 spine level (left). SpEps, either by T11 stimulation with T3 recording or vice versa, showed similar waveforms with the same NI and NII latencies. [Reproduced with permission from Machida et al. (5).]

used for surgeries of the thoracic or lumbar spinal cord.

Disadvantages of Epidural Spinal Cord Stimulation and Recording Although there are many advantages of epidural spinal cord stimulation and recording, there are some disadvantages as well, which limit their use. The main disadvantage is the insecurity of electrode placement; the electrodes can easily be dislodged. Unless the patient is paralyzed, even a low-intensity stimulus will cause vigorous paraspinal muscle twitches that could interfere with the operative procedure. The epidural electrodes can easily be applied when the spine is exposed, as in some neurosurgical procedures; however, for others, such as those with an endovascular

approach, the epidural electrodes must be inserted percutaneously. Insertion of electrodes in the epidural space, although uncommon, poses the potential risk of injuring the spinal cord.

Motor Evoked Potentials In order to evaluate the anterior spinal cord or motor function, it is necessary to monitor descending motor pathways. This also can be done by several methods. The stimuli can be applied transcranially by electrical or magnetic stimulation or to the spinal cord via epidural electrodes. The recording may be from the spinal cord, peripheral nerves, or muscles. Although transcranial magnetic stimulation is effective and painless

Husain 08

1/17/08

11:59 AM

Page 125

CH A P T ER 8: Spinal Cord Surge r y

in an awake subject, it is not suitable for NIOM because it is highly sensitive to the orientation of the stimulating coil as well as to anesthetics. Transcranial magnetic stimulation is not used for NIOM and is not discussed here. The most commonly used method is transcranial electric stimulation (TES) and recording compound muscle action potentials (CMAPs) from muscles. This is described further on. In this section, other methods of MEP recording are described.

Transcranial Electric Stimulation and Recording from the Spinal Cord The potential is recorded from the caudal spinal cord via an epidural electrode and consists of a bi- or triphasic sharp discharge, called a D (direct) wave, followed by a series of polyphasic waves, called I (indirect) waves (5,6) (Figure 8.3). Stimulation is delivered to the scalp with an anode at the vertex and the cathode 6 to 7 cm anterior or lateral to the vertex. The stimulus intensity is usually 400 to 500 V of 50 to 100 µs duration, which gives approximately 400 to 600 mA in current intensity. The D wave results from direct stimulation of corticospinal neurons, whereas I waves are generated by transsynaptic activa-

•

1 25

tion of corticospinal neurons. D waves are resistant to inhalation anesthetics but I waves are extremely sensitive. Since the D wave’s amplitude is robust (15 to 25 µV at high- thoracic and 5 to 10 µV at low-thoracic levels), a single stimulus usually yields a measurable response, but 15 to 25 sweeps may be averaged for a betterdefined response. The major advantage of this method is that neuromuscular blocking agents and inhalation anesthetics can be used without affecting the response.

Spinal Cord Stimulation and Recording from Muscles or Nerves A low-intensity stimulus (less than 10 mA) applied to the rostral thoracic spinal cord via epidural electrodes can elicit CMAPs from lower extremity muscles and nerve action potentials (NAPs) from the tibial and peroneal nerves (7,8) (Figure 8.4). CMAPs are robust responses and no averaging is necessary to elicit a measurable response. NAPs can be recorded by averaging 10 to 25 sweeps. The advantage of both techniques is their insensitivity to anesthetics. The use of paralytic agents, however, attenuates or abolishes the CMAPs but not the NAPs. However, both CMAPs and NAPs elicited in this manner are likely not to reflect motor function of the spinal cord. Rather, they are antidromic sensory responses or mixed motor and sensory responses. Regardless, this method monitors at least certain functions of the spinal cord and can be useful in some cases when anesthetics or neuromuscular blocking agents cause problems in eliciting measurable responses.

Common NIOM Paradigm for Spinal Cord Surgery FIGURE 8.3 D and I waves recorded from the spinal cord via epidural electrodes after TES. Stimulus intensity 100% = 750 V. Note that only the D wave is elicited by a weak stimulus and the latency shortens with the greater stimulus intensity. [Reproduced with permission from Deletis (6).]

NIOM during spinal cord surgery usually involves SEP and MEP monitoring. SEPs are usually obtained with peripheral nerve stimulation and recorded with scalp electrodes. Stimulation for MEPs is at the scalp with

Husain 08

126

1/17/08

•

11:59 AM

Page 126

SE C T I ON I I : C l i n i c a l Me thod s

FIGURE 8.4 Simultaneously recorded tibial NAPs (upper three tracings) and anterior tibialis CMAP (lower three tracings) after stimulation of the spinal cord via epidural electrodes placed at the T2 spine level.

CMAP recordings from muscles. At times, the recording of D waves may be useful as well. The methodology used in obtaining these responses in the authors’ practice is discussed below.

Somatic Evoked Potentials SEP recording montages may differ depending on the preference of the laboratory. Subcortical far-field potentials (P14-N18 for upper and P31-N34 for lower extremities) are important to record, as they are affected little by anesthesia and consequently are a more consistent and reliable potential than the cortical potentials (N20 for upper and P37 (P40) for lower extremities). These far-field potentials can be recorded from any scalp electrode (the exact scalp location is not critical for farfield potential recordings because of a wide field distribution) using a noncephalic refer-

ence. Instead of using a distant reference, which may be prone to contamination by the electrocardiogram (ECG) or other physiologic or nonphysiologic artifacts, ear reference recordings are technically much easier and yield cleaner far-field potentials. Commonly used montages for upper and lower extremity SEPs are presented in Tables 8.2 and 8.3. The authors primarily focus on and follow far-field potentials (P14 for upper and P31 for lower extremities) of the SEPs (Figure 8.5). Channels 1 and 2 are redundant, but recording both helps in cases where one does not work. Channel 3 records only the cortical potential and is the least useful, but this may help in case the far-field recording fails. Monitoring peripheral potentials (channel 4) is important because this assures the appropriate delivery of stimulus. One average requires about 500 to 1,000 responses, which usually takes several min-

Husain 08

1/17/08

11:59 AM

Page 127

CH A P T ER 8: Spinal Cord Surge r y TABLE 8.2 A Typical Montage Used for Upper Extremity SEPs

Channel

Montage

Waveform(s) recorded

•

1 27

TABLE 8.3 A typical Montage Used for Lower Extremity SEPs

Channel

Montage

Waveform(s) recorded

1

CP3/4 (i)—A1 + A2

P14—N20

1

CPz—A1+A2

P31—P37

2

CP3/4 (c)—A1 + A2

P14—N18

2

CP1/2 (i) - A1+A2

P31—P37

3

CP3/4 (c)—CP3/4 (i)

N20

3

CPz—Fpz

P37

4

Erb (i)—Erb (c)

Erb potential

4

Popliteal fossa

Popliteal fossa potential

(i) = ipsilateral to side of stimulation; (c) = contralateral to side of stimulation.

(i) = ipsilateral to side of stimulation.

FIGURE 8.5 Typical examples of 16-channel recording of left/right ulnar and tibial nerve SEPs based on Tables 8.2 and 8.3 montages arrangement (only 8 channels are shown here). Each stimulus was delivered sequentially starting with the left and right ulnar nerves and then left and right tibial nerves with a slight latency delay (50 ms each). Note that the P14 of the ulnar and P31 of the tibial nerve SEPs are registered only with an ear reference but not with a F7 reference.

Husain 08

128

1/17/08

•

11:59 AM

Page 128

SE C T I ON I I : C l i n i c a l Me thod s

utes with a stimulus rate of 5 to 7 Hz. A caveat to this monitoring method is when total intravenous anesthesia (TIVA) is used. In this situation, cortical responses are robust and easily reproducible, making subcortical responses less important. When surgery is performed on the caudal part of the cervical spinal cord or lower, tibial nerve SEPs are continuously averaged. Median nerve SEPs can serve as controls in such cases and are performed after several tibial nerve SEP averages. However, if the pathology is in the rostral cervical spinal cord, median or ulnar nerve SEPs should be used primarily. Ulnar nerve SEPs, though harder to obtain, provide more coverage of the cervical spinal cord. In such surgeries, tibial nerve SEPs should be obtained as redundant information at periodic intervals.

Motor Evoked Potentials The stimulating electrodes for MEPs are placed 2 cm anterior to C3 (C3'), C4 (C4'), Cz, (Cz'), C1, C2, and Fz. These may be surface disc, needle, or corkscrew electrodes. Low electrode impedance for both recording and stimulating electrodes is important, which minimizes the stimulus artifact and decrease the stimulus intensity required to elicit reliable responses (Figure 8.6). The stimulating electrodes must be covered with a nonconductor to avoid delivering accidental shock to the operating room personnel (especially the anesthesiologist), other instruments, or surgical devices. Using C3' and C4' stimulus sites usually elicits MEPs in both upper and lower extremities, although Cz'-Fz or C1-C2 may be used for lower

FIGURE 8.6 MEPs from left/right abductor pollicis brevis, brachioradialis, tibialis anterior, and abductor hallucis muscles after TES. Stimulating electrodes were placed at C3 and C4. In this case MEPs were recorded bilaterally with either-side stimulation. In this figure greater amplitude is noted in the right-sided muscles with stimulation of the left hemisphere by anodal electrode at C3.

Husain 08

1/17/08

11:59 AM

Page 129

CH A P T ER 8: Spinal Cord Surge r y

extremity MEPs. The best sites for the stimulation of upper and lower MEP recordings are variable; in some cases changing the stimulus sites may improve responses and decrease the stimulus artifact. Unlike peripheral nerve stimulation, in which the cathode acts as the stimulus site, in TES the anode is the site for excitation. Therefore if the anode is on the left side of the scalp, this primarily activates the left hemisphere, resulting in activation of right limb muscles. The TES to elicit MEPs requires the stimulating device to be capable of delivering high voltage (up to 1,000 V) and multipulse stimulation (as with the Digitimer). Multipulse stimulation is more effective in yielding a robust and consistent MEP than single-pulse stimulation due to temporal summation. Usually a train of three to five stimuli with an interstimulus interval of 1 to 3 ms is used. The stimulus intensity and duration parameters have been noted above. Whether MEPs are best recorded from muscle (CMAPs), spinal cord (D and I waves), or both is still in dispute. Generally the spinal cord recording is more popular in Europe, Japan, and Australia, but in the United States most laboratories use CMAP recordings. Combining CMAP and D-wave MEP recordings has been shown to be of particular utility in spinal cord surgery, as discussed further on (9). Commonly used muscles for recording CMAPs include the brachioradialis and abductor pollicis brevis muscles in the upper limb and anterior tibialis and abductor hallucis brevis muscles in the lower limb. Small muscles generally produce robust MEPs because of their rich innervation by corticospinal fibers. Ideally the first trial of MEP recording is done shortly after the patient is anesthetized and before the start of electrocautery use. However, sometimes the effect of neuromuscular blocking drugs used during induction may prevent obtaining CMAP MEPs until later. Once the electrocautery starts, it is almost impossible to get a good MEP base-

•

1 29

line. Warning the surgeon and surrounding personnel before delivering a MEP stimulus is important, so that sudden patient movements will not disrupt the surgery. The first MEP stimulus should start with 0 V to verify that the stimulus correctly triggers the sweep. The stimulus intensity is then increased to 100 V, which should give about 100 mA or greater current intensity. If the current delivered is extremely low, the impedance of the stimulating electrodes should be checked. If no response is elicited, the intensity is increased in 50- to 100-V increments until an adequate MEP is obtained. MEPs should be performed every 2 to 5 minutes. Even more frequent trials can be obtained during dissection of the spinal cord lesion. If D waves are to be recorded, an epidural electrode must be placed caudal to the tumor. This can be placed percutaneously by the surgeon, anesthesiologist, or neurophysiologist. Alternatively, once the laminectomy has been performed, the surgeon can place the electrode caudal to the surgical site and suture it in place. When D waves are to be recorded, only a single TES is needed. However, the stimulus intensity should be increased as described above until a reliable response is recorded. It is best to alternate between SEPs and MEPs; however, some surgeons prefer to tell the NIOM technologist when to stimulate for MEPs. This should be clarified with the surgeon prior to the surgery.

ANESTHETIC CONSIDERATIONS Anesthesia can be beneficial as well as detrimental for NIOM during spinal cord surgery. When the patient is anesthetized, a higher stimulus intensity can be used than in an awake state. Anesthetics also reduce movement and muscle artifacts, yielding better and “cleaner” responses than recorded in an awake state. However, anesthesia also changes SEP and MEP waveforms consider-

Husain 08

130

1/17/08

•

11:59 AM

Page 130

SE C T I ON I I : C l i n i c a l Me thod s

ably, and the responses under anesthesia are quite different from those obtained while the patient is awake. All inhalation agents attenuate or abolish CMAP MEPs and prolong the latencies. CMAP MEPs are more sensitive to anesthesia than SEPs.

Inhalation Anesthesia The common inhalation anesthetics are nitrous oxide (N2O), halothane, desflurane, isoflurane, etc. The relative potency of inhalation agents is described by the minimal alveolar concentration (MAC), which is the anesthetic concentration needed to prevent movement to a painful or surgical stimulus in 50% of patients. The actual concentration needed to obtain one MAC varies among anesthetics. For example, the MAC for isoflurane is 1.3% and that for halothane is 0.7%. The short-latency oligosynaptic SEP components are more resistance than long-latency polysynaptic components. Thus, short-latency far-field evoked potentials (P14 and P31 for the upper and lower extremities, respectively) are affected little with inhalation anesthesia. Often one can obtain consistent and welldefined far-field SEPs with isoflurane greater than 1% with 50% to 60% N2O. However, CMAP MEPs usually cannot be recorded with isoflurane greater than 0.5%. D waves are little affected by inhalation anesthetics.

seizures, especially in patients with a history of seizures. Most often, when both SEPs and MEPs must be recorded, TIVA with propofol and an opioid is used.

Neuromuscular Blocking Drugs Neuromuscular blocking drugs (pancuronium, vecuronium, atracurium, etc.) are often used in combination with intravenous or inhalation anesthetics. “Balanced anesthesia” includes narcotics for analgesia plus a neuromuscular blocking drug with a relatively weak anesthetic, such as N2O. The neuromuscular blocking drugs do not affect SEPs. However, they eliminate muscle twitches in association with stimulation of a mixed nerve. Therefore they may give the erroneous impression that an adequate stimulus was not delivered to the nerve. The NIOM technologist should document the degree of paralysis if neuromuscular blocking drug are used. Their use is detrimental for CMAP MEPs, but the latter can be recorded if there are two twitches in a train-of-four repetitive stimulation. D waves are not affected by neuromuscular blocking drugs. If it is necessary for the patient to have complete neuromuscular blockade, D waves should be used for monitoring the MEPs.

WARNING CRITERIA Total Intravenous Anesthesia TIVA has recently become increasingly popular for NIOM. TIVA includes barbiturates, opioids (morphine, alfentanil, sufentanil, remifentanil, etc.), propofol, etomidate, and ketamine. The effect of barbiturates on evoked potentials is similar to that of halogenated agents. Propofol has a short duration of action and little effect on SEPs and MEPs unless a very high dose is used. Both ketamine and etomidate may enhance evoked potential responses by heightening synaptic function at a low dose. However, both may induce

The warning criteria for SEPs have arbitrarily been accepted as greater than 50% reduction of amplitude and/or a 10% prolongation of latency of a given component. There is little consensus regarding the warning criteria for CMAP MEPs. Some consider a greater than 50% amplitude reduction of CMAPs as significant; however, because of their variability, others consider only a complete or near complete loss of response as significant. In some situations, abrupt loss or marked attenuation of CMAP MEPs can occur right after recording a robust response. The reason for

Husain 08

1/17/08

11:59 AM

Page 131

CH A P T ER 8: Spinal Cord Surge r y

this marked fluctuation of amplitude is not known but may be involve the spontaneous fluctuation of excitability of the anterior horn cells. The loss or attenuation of CMAP MEPs, therefore, must be verified before deciding that the change is “real.” Warning criteria for D waves may be more reliable than those for CMAP MEPs, as D waves are less sensitive to systemic and anesthetic changes; therefore marked fluctuations in amplitude are uncommon. When corticospinal tract fibers are damaged, there is an incremental decrease in the D-wave amplitude, and a decrease of 50% is considered significant (9). Drops of 30% to 50% are considered minor warning signs, and the surgeon may chose to pause the surgery transiently when this occurs. A drop of 50% or more, however, should discourage the surgeon from further dissection. A sudden drop in Dwave amplitude indicates vascular compromise and is often irreversible. Combining CMAP and D-wave MEP monitoring has been noted to be particularly useful in spinal cord surgery by some investigators (9–11). Preservation of both CMAP and D-wave MEPs during surgery is consistent with no lower extremity weakness postoperatively. Loss of CMAP MEPs with preservation of D waves (amplitude greater than 50% of baseline) suggests transient postoperative paraparesis. Loss of both CMAP and D-wave MEPs will result in persistent weakness of the lower extremities. Thus, if there is loss of the CMAP MEPs and D waves are not being recorded during removal of a spinal cord tumor, the surgeon will likely stop dissecting the tumor and perform an incomplete resection. If, however, D waves are being recorded and their amplitude is maintained when CMAP MEPs are lost, the surgeon can continue dissection, thereby allowing a more complete resection. A postoperative paraplegia will likely occur but will be transient. It should be cautioned that this methodology has been tested in only a few centers.

•

1 31

After the NIOM team has alerted the surgeons of changes in responses, the surgeons must decide if they want to modify the surgical procedure, continue with the case, or wait until the responses recover. Data from a combination of SEP, CMAP, and D-wave MEP monitoring can help them considerably in making their decision. In some cases, either only SEPs or MEPs can be recorded or responses are small and “noisy.” In such cases the surgeon must be made aware that recordings are suboptimal and may not be reliable. When a posterior myelotomy is necessary, as in surgery for intramedullary tumors, the injury to the spinal cord is mostly mechanical and inflicted preferentially to the dorsal columns. With a posterior myelotomy, it is expected that SEPs transmitted through the dorsal columns will become impaired (Figure 8.7A, B, and C). Under those circumstances, the corticospinal tracts are spared from injury, and MEPs may be preserved. In such cases, MEPs should be followed more closely. Excision of extramedullary tumors—such as meningiomas, nerve sheath tumors, and metastases—often requires some traction of the spinal cord. Overzealous retraction of the spinal cord can result in a unilateral mechanical injury. Such an injury to the spinal cord can occur wherever the traction is applied. Depending on the location of the traction, changes in SEPs or MEPs may be noted. To facilitate exposure and tumor resection, especially if the tumor is ventral, the surgeon may resort to nerve root sectioning. Sacrificing a nerve root in the thoracic spinal cord is usually not associated with any significant neurologic deficit, since many of the nerves overlap in their dermatomal distribution. However, radicular arteries that supply the spinal cord travel within the dural sheaths of the nerve roots. Thus when a nerve is sacrificed in the neural foramen, the radicular artery is also sacrificed. If this radicular artery is a major contributor to spinal cord blood supply, such as the ARM, or if several roots are sacrificed, spinal cord ischemia and infarc-

Husain 08

132

1/17/08

•

11:59 AM

Page 132

SE C T I ON I I : C l i n i c a l Me thod s

A

B

FIGURE 8.7 A. Median nerve SEP monitoring during surgery for intramedullary cervical spinal cord tumor. After a midline myelotomy, there was sudden lost of SEP to right-side stimulation. B. The preoperative median nerve SEP was normal. C. The postoperative SEP showed absence of the P14 and subsequent peaks to right-side stimulation. Postoperatively, the patient had severe loss of proprioceptive sensation of the right arm.

Husain 08

1/17/08

11:59 AM

Page 133

CH A P T ER 8: Spinal Cord Surge r y

•

1 33

C

tion can occur. The thoracic spinal cord is most vulnerable to ischemic damage because this region is considered a watershed zone with relatively sparse collaterals. This type of ischemia would affect the MEPs more significantly than the SEPs, as the ventral spinal cord would be affected more. Sectioning of nerve roots is therefore ill advised.

UTILITY OF NIOM There is a growing body of literature suggesting that NIOM can reduce morbidity in spinal cord surgeries. Until recently, most data were for SEP monitoring. In one series, 20 patients were monitored with SEPs; 6 of these developed postoperative deficits. In one patient, baseline SEPs were not obtained in the limb that developed weakness. In all the others, SEP changes occurred during dissection, warning of postoperative complications (12). In another series of 30 patients undergoing spinal cord surgery for pain syndrome, intramedullary tumors, and syringomyelia,

changes in subcortical SEP waveforms suggested damage to the spinal cord; as a result, the surgical procedures in several cases were modified (13). There are no reports in which SEP monitoring did not detect impending spinal cord damage. However, in other types of surgeries on the spine, preservation of SEPs has been noted despite the occurrence of postoperative paraplegia; this most likely happens with spinal cord surgery as well. In the last decade, the utility of MEP monitoring has been reported. It should be noted that SEP monitoring is almost always performed in addition to MEP monitoring. In a series of 28 patients who underwent CMAP MEP monitoring after TES, significant changes were noted in 13 patients; 12 of these had persistent postoperative paraparesis (14). In another series of 32 patients, D-wave MEP monitoring was possible in 19 (15). A decrease in D-wave amplitude of greater than 50% was seen in 3 patients; all 3 had postoperative paraplegia, which eventually recovered in 2. Postoperative weakness was not noted in any of the patients who had a decrease in D-

Husain 08

134

1/17/08

•

11:59 AM

Page 134

SE C T I ON I I : C l i n i c a l Me thod s

wave amplitude of less than 50%. Of the 13 patients in this series in whom MEP monitoring was not possible, 5 had postoperative weakness. The most convincing data on the utility of NIOM come from a recent historical control study (16). In this study, outcomes of 50 patients who underwent NIOM with SEPs, CMAPs, and D-wave MEPs were compared to matched controls who had previously had surgery without NIOM. The investigators used the SEP and MEP interpretation criteria described earlier. The patients who had NIOM were less likely to experience a deterioration in neurologic status as compared to controls.

TECHNICAL CONSIDERATIONS Preparation One of the main responsibilities of the NIOM technologist is to communicate with the patient and his or her family about the NIOM and how this will be achieved. The patient should be assured that NIOM usually does not hurt, but if acetone and/or collodion is used to attach the electrodes, there may be a strong associated odor. Many patients presenting for spinal cord surgery will have neurologic deficits caused by the spinal cord lesion. These deficits, such as lower extremity weakness, should be assessed, as they may affect MEP NIOM. Other medical problems, such as a peripheral neuropathy and orthopedic problems, should also be noted, as they may affect SEP monitoring. The technologist should also determine if there are any relative contraindications for MEP monitoring, such as the presence of a cardiac pacemaker, other implantable devices, seizures, etc. This information must be conveyed to the neurophysiologist, and, with the surgeon’s input, the neurophysiologist and technologist should decide which modalities of monitoring are most appropriate for the scheduled surgery.

Preoperative baseline SEP studies should be obtained whenever possible. Because the preoperative baseline studies are done in the awake state, the patient may not able to tolerate the sufficient stimulus intensity or may be unable to relax to yield “clean” responses. It is not uncommon to obtain well-defined responses during surgery when preoperative baseline responses were poorly reproducible. If baseline responses are not present, the NIOM team should be prepared to perform the more unusual types of monitoring discussed previously. The NIOM technologist should also ensure that he or she has all the necessary electrodes for the surgery. If an epidural electrode will be used to record D waves, arrangements must be made with the surgeon, anesthesiologist, or neurophysiologist about implanting it. Almost any modern NIOM machine can be used for these surgeries; however, having at least a 16 channels available is important. This will allow performing upper and lower limb SEPs and CMAP MEPs from all four extremities, as well as D waves if necessary. The NIOM machine should be capable of delivering a high-intensity TES; if it is not, a separate machine for this purpose should be used, such as the Digitimer.

Procedure Electrode application on the day of the surgery can be done in the preoperative holding area. Needle electrodes should not be applied when the patient is awake. Instead, once the patient is anesthetized, these electrodes can easily be inserted. In special situations where the patient is extremely frightened or cognitively impaired, all the electrodes should be applied after the patient has been anesthetized. There are many types of recording and stimulating electrodes available in today’s market. Whichever electrodes are chosen, it is important to maintain consistency in application and placement. Skin preparation should

Husain 08

1/17/08

11:59 AM

Page 135

CH A P T ER 8: Spinal Cord Surge r y

be done with a light sandpaper or skin preparation product such as Nu Prep or Lemon Prep. This should be done for surface and selfadhesive electrodes. Placement of scalp electrodes should be based on the international 10-20 system. It is also wise to place redundant electrodes in case some electrodes do not work or become disconnected. If the epidural electrode is to be inserted percutaneously, it should be done before the patient is draped. Alternatively, if it will be placed after the laminectomy, it should be made available to the scrub nurse to pass to the surgeon at the appropriate time. It is ideal if the first sets of SEPs and MEPs are obtained before the patient is draped and covered in the sterile field. This will establish the integrity of the recorded system and allows adjustment of stimulating or recording electrodes if necessary. This is an important step, as once the patient is draped, it is virtually impossible to change or reapply the electrodes. These steps will ensure successful recording after the patient is placed under the sterile field. The first operative baseline studies are done after the patient is in a stable anesthetic state. This is the “true” baseline response and the one to which any changes in the operating room must be compared. This baseline is likely to be different than the preoperative baseline study obtained in the outpatient laboratory, as the patient was awake during the latter study. Under anesthesia, unlike the SEPs recorded in the awake state, middle- and longlatency components are abolished and only the short-latency components are preserved. In some cases, especially with high doses of inhalation anesthetics, even the first cortical potential (N20 for upper and P37 for lower SEPs) may not be present. For surgeries involving the upper cervical spine, median or ulnar nerve SEPs and CMAP MEPs from all four extremities should be monitored. For lower spinal cord lesions, tibial nerve SEPs and CMAP MEPs should be obtained. For lower spinal cord surgeries,

•

1 35

median SEPs should also be obtained and serve as a control. Recording D waves from an epidural electrode should also be considered. Alternating between SEP and MEP monitoring is useful except if a posterior myelotomy is performed; this renders SEP monitoring impossible; thus MEP monitoring becomes more important. During tumor or lesion dissection, frequent MEPs should be obtained. The surgeon should always be warned prior to the delivery of a MEP stimulus so that sudden patient movement does not disrupt the surgery. A crucial role of the NIOM technologist is to recognize and troubleshoot artifacts. The operating room is an electrically hostile environment. There can be several types of interference and artifacts causing problems for NIOM. The blood warmers, warming blankets, electrocautery, CUSA, x-ray machine, surgical microscope, and anesthesia machine may give off very high frequency noise. The technologist should be able to identify the different types of artifact and adjust accordingly. Sometimes it can be as simple as regelling an electrode or unplugging the device that is causing the problem. Operating room equipment should never be unplugged without first conferring with the circulating nurse. At other times, the artifact can be recalcitrant and difficult to eliminate. Troubleshooting can best be accomplished by using a systematic and organized approach. If a significant change in responses is observed, the NIOM team must determine the source, so that appropriate corrective steps can be taken. The first step is to make certain that the change is reliable and consistent, not just a one-time instance. At least two sets of responses should be obtained to verify change. If the change is verified and cannot be attributed to systemic processes or artifact, it must be localized to the brain, cervical spinal cord, or peripheral nervous system. The significance of the change should be interpreted by the neurophysiologist, and the surgeon should be alerted.

Husain 08

136

1/17/08

•

11:59 AM

Page 136

SE C T I ON I I : C l i n i c a l Me thod s

Several steps must be taken to troubleshoot a decrement in NIOM responses. The first thing to check is the original incoming raw data. If one channel has a lot of artifact, it may be due to a high impedance or loss of an electrode. If all channels are affected, there could be loss of the ground electrode or introduction of an external interference. With newer equipment, the technologist can monitor original data continuously throughout the procedure together with averaged responses. It helps to determine whether the incoming raw EEG data are intermittently contaminated by the external physical or electrical interference. Next, an impedance check must be performed. Thereafter, the stimulators should be checked to ensure that adequate stimulation is being given. The right and left stimulators can be switched as a test. If the problem moves to the other side, the stimulator cable is most likely at fault. Finally, sources of external interference should be sought. A new instrument can cause electrical noise, a patient under very light anesthesia will have EMG contaminating the average, an unexpected bolus of medicine given by the anesthesiologist may cause a drop in amplitude of the response, and changes in temperature and blood pressure can also have significant effects. Occasionally situations are encountered in which 60-Hz or other high-frequency external interference or large stimulus artifact cannot be eliminated. Recording parameters such as frequency bandwidth, averaging number, or stimulus rate can be changed to help eliminate the artifact. The high-frequency filter may be reduced to 500 Hz or less and the low filter may be increased to 20 to 50 Hz. With a filter setting change, waveform latency and amplitude may change. Use of a 60-Hz filter can produce “60-Hz ringing,” which consists of rhythmic waves after the stimulus artifact, mimicking the “reproducible evoked response.” If the external interference is extremely large and most trials are rejected, it

may be necessary to decrease the sensitivity or change the rejection threshold.

Postprocedure NIOM for most spinal cord surgeries continues until the surgical wound is closed. Electrodes are removed at the end of the case. Needle electrodes are removed slowly, so as to not to cause injury. Disposable electrodes are discarded in an appropriate container. Any reusable electrodes, such as surface disk electrodes, are placed in a plastic bag, later to be cleaned and disinfected according to laboratory protocol. If possible, the NIOM technologist should examine the patient after his or her recovery from anesthesia to document adequate movement and sensation in the limbs. All the documentation is completed with a copy of all the waveforms and log of events printed for the report. In addition to the recording parameters, all alerts issued to the surgeon should be well documented, along with any changes that occurred owing to the alerts. The neurophysiologist should create a report documenting all these findings.

CONCLUSIONS Spinal cord surgery can be associated with high neurologic morbidity, paraparesis being the most significant complication. NIOM offers surgeons a technique that can help reduce the incidence of postoperative paraparesis. Unlike other types of spinal surgery, surgery on the spinal cord itself is likely to result in selective damage to various fiber tracts, making multimodality monitoring essential. Advances in MEP monitoring have helped reduce morbidity and may be more specific in predicting impending permanent postoperative paraparesis. NIOM for spinal cord surgery should include at least SEP and CMAP MEP monitoring and preferably Dwave monitoring as well.

Husain 08

1/17/08

11:59 AM

Page 137

CH A P T ER 8: Spinal Cord Surge r y

REFERENCES 1. Lesser RP, Raudzens P, Luders H, et al. Postoperative neurological deficits may occur despite unchanged intraoperative somatosensory evoked potentials. Ann Neurol 1986;19:22–25. 2. Zornow MH, Grafe MR, Tybor C, Swenson MR. Preservation of evoked potentials in a case of anterior spinal artery syndrome. Electroencephalogr Clin Neurophysiol 1990; 77:137–139. 3. Sansur CA, Pouratian N, Dumont AS, et al. Part II: spinal-cord neoplasms: primary tumours of the bony spine and adjacent soft tissues. Lancet Oncol 2007;8:137–147. 4. Traul DE, Shaffrey ME, Schiff D. Part I: spinal-cord neoplasms: intradural neoplasms. Lancet Oncol 2007;8:35–45. 5. Machida M, Weinstein SL, Yamada T, Kimura J. Spinal cord monitoring. Electrophysiological measures of sensory and motor function during spinal surgery. Spine 1985;10: 407–413. 6. Deletis V. Intraoperative neurophysiology and methodologies used to monitor the functional integrity of the motor system. In: Deletis V, Shils J, eds. Neurophysiology in Neurosurgery: A Modern Intraoperative Approach. San Diego, CA: Academic Press, 2002:25–51. 7. Erwin CW, Erwin AC. Up and down the spinal cord: intraoperative monitoring of sensory and motor spinal cord pathways. J Clin Neurophysiol 1993;10:425–436. 8. Owen JH. The application of intraoperative monitoring during surgery for spinal deformity. Spine 1999;24:2649–2662. 9. Kothbauer K, Deletis V, Epstein FJ. Intraoperative spinal cord monitoring for

10.

11.

12.

13.

14.

15.

16.

•

1 37

intramedullary surgery: an essential adjunct. Pediatr Neurosurg 1997;26:247–254. Deletis V. Neuromonitoring. In: McLone DG, ed. Pediatric Neurosurgery: Surgery of the Developing Nervous System. Philadelphia: Saunders, 2001:1204–1213. Sala F, Lanteri P, Bricolo A. Motor evoked potential monitoring for spinal cord and brain stem surgery. Adv Tech Stand Neurosurg 2004;29:133–169. Kearse LA Jr, Lopez-Bresnahan M, McPeck K, Tambe V. Loss of somatosensory evoked potentials during intramedullary spinal cord surgery predicts postoperative neurologic deficits in motor function [corrected]. J Clin Anesth 1993;5:392–398. Prestor B, Golob P. Intra-operative spinal cord neuromonitoring in patients operated on for intramedullary tumors and syringomyelia. Neurol Res 1999;21:125–129. Quinones-Hinojosa A, Lyon R, Zada G, et al. Changes in transcranial motor evoked potentials during intramedullary spinal cord tumor resection correlate with postoperative motor function. Neurosurgery 2005;56:982–993; discussion 982–993. Morota N, Deletis V, Constantini S, et al. The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery 1997;41:1327–1336. Sala F, Palandri G, Basso E, et al. Motor evoked potential monitoring improves outcome after surgery for intramedullary spinal cord tumors: a historical control study. Neurosurgery 2006;58:1129–1143; discussion 1129–1143.

Husain 08

1/17/08

11:59 AM

Page 138

Husain 09

1/17/08

12:00 PM

Page 139

9

Lumbosacral Surgery

Neil R. Holland

NERVE ROOT MONITORING USING FREE-RUNNING EMG

T

he adult spinal cord ends opposite the lower border of the L1 vertebra; below that, the spinal canal contains only individual lumbosacral nerve roots descending toward their respective neural foraminae, collectively known as the cauda equina. Nerve roots can be injured during lumbosacral surgery by probing, drilling, misplaced hardware, or stretching during correction of spinal deformity. There is a 10% risk of new postoperative neurologic deficit from nerve root injury during spinal fusion for degenerative lumbosacral spine disease, 10-fold greater than the risk of spinal cord injury from scoliosis surgery. The most common deficit is a new foot drop from surgical injury to the L5 nerve root, although findings can include any pattern of dermatomal numbness, radicular weakness, and/or sphincter dysfunction. Nerve root injuries are more frequent during revision than primary surgery and in cases where multiple levels are fused (1). Obviously, the purpose of neurophysiologic intraoperative monitoring (NIOM) with electromyography (EMG) during these lumbosacral spine cases is to detect early and potentially reversible surgical nerve root irritation, alert the surgeon, and thus prevent more significant injury and postoperative deficit.

Background Mixed nerve somatosensory evoked potentials (SEPs) and transcranial electrical motor evoked potentials (MEPs) are the NIOM techniques most often used for spinal cord monitoring, but they are not sensitive for detecting nerve root injuries. SEPs represent the summation of neural signals that enter the spinal cord through multiple segments, and because of central amplification, they can remain completely unchanged after an individual nerve root lesion. MEPs recorded from an epidural electrode will obviously not be affected by a more caudal nerve root injury. MEPs recorded from a single distal limb muscle throughout surgery will not detect nerve root injuries affecting other myotomes. There have been numerous well-documented cases of patients with acute radiculopathy from spine surgery despite normal SEP and MEP monitoring. Dermatomal somatosensory evoked potentials (DSEPs) can detect individual nerve root injuries but are technically difficult to resolve in the operating room; moreover, they need to be averaged many times, resulting a slow turnaround time, and

139

Husain 09

140

1/17/08

•

12:00 PM

Page 140

SE C T I ON I I : C l i n i c a l Me thod s

they test only dorsal root (sensory) function. For these reasons EMG has become the neurophysiologic technique most often used to monitor nerve root function during lumbosacral spine surgery. EMG is more sensitive than SEPs and DSEPs for detecting nerve root dysfunction, provides immediate results without averaging, and can be monitored from multiple channels (i.e., multiple myotomes and nerve roots) simultaneously, giving instantaneous feedback to the surgeon during critical phases of the procedure.

Technique Continuous free-running EMG is monitored from muscles innervated by nerve roots considered to be at risk for injury during surgery. Blunt mechanical trauma to nerve roots causes a depolarization that will be conducted

down the nerve, across the neuromuscular junction, and evoke identifiable motor unit potentials in the monitored muscle (Figure 9.1). Minor nerve root manipulation will result in a short burst of motor unit potentials (Figure 9.2A). More severe mechanical nerve root injury or retraction will cause prolonged trains of high-frequency motor unit potentials, often referred to as neurotonic discharges (Figure 9.2B and C). Repetitive grouped motor unit potentials, myokymic discharges, are seen less frequently and usually indicate very severe nerve injury (Figure 9.2D). These are the very same EMG abnormalities seen in the diagnostic laboratory in cases of spontaneous motor nerve hyperexcitability (Isaac’s syndrome). Identification of this abnormal EMG activity can be used to alert the surgeon of inadvertent trauma to nerve roots, allowing evasive action in an

FIGURE 9.1 Free-running EMG activity for nerve root monitoring. A. EMG monitoring should be quiescent under normal conditions. B. Blunt mechanical nerve root irritation activates the motor nerve fibers, is transmitted down the nerve and across the neuromuscular junction, and evokes recordable motor unit potentials in the monitored muscle.

Husain 09

1/17/08

12:00 PM

Page 141

CH A P T ER 9: Lumbos acral Surge r y

A

B

C

D

•

1 41

FIGURE 9.2 EMG activity from nerve root irritation or injury. A. Minor nerve root manipulation evokes a short burst of motor unit potentials. B and C. More severe blunt mechanical nerve root injury or retraction evokes prolonged trains of high-frequency motor unit potentials, known as neurotonic discharges. D. Rarely, myokymic discharges can be seen, but usually only after very severe nerve injury. In each case the time base shown is 100 ms/div and the display sensitivity is 50 µV/div.

effort to prevent more severe or irreversible injury. The loudspeaker on the EMG machine can be used to provide the surgeon with an instantaneous audible warning of potential nerve root injury. Limb muscles can be used to monitor spinal levels down to the S2 nerve roots. The anal and/or urethral sphincter muscles should be monitored in cases where the conus medullaris or lower sacral nerve root are felt to be at particular risk for iatrogenic injury (Table 9.1).

sensitivity, but a higher specificity and positive predictive value (2) (Table 9.3). However, sensitivity is preferable to specificity for NIOM, as it is more important to detect a mild degree of nerve root irritation and alert the surgeon in time to take evasive action, than it is to correctly predict an irreversible postoperative deficit. TABLE 9.1 Muscles Suitable for Intraoperative Electromyographic Monitoring of Thoracolumbosacral Nerve Root Segments during Lumbar Fusion

Outcome Data Significant EMG activity occurs in 80% of monitored lumbosacral spine surgeries, although less that 10% of these develop a new persistent postoperative neurologic deficit (Table 9.2). In other words, EMG monitoring has a high sensitivity for detecting mechanical nerve root manipulation but a low specificity and positive predictive value for postoperative nerve root injury. SEP monitoring has a lower

Nerve roots at risk

Appropriate muscle for EMG monitoring

T7–12

External oblique and rectus abdominus

L1–2

Iliacus

L2–4

Vastus medialis

L4–5

Tibialis anterior

S1–2

Medial gastrocnemius

S3–5

Anal and urethral sphincter

Husain 09

142

1/17/08

•

12:00 PM

Page 142

SE C T I ON I I : C l i n i c a l Me thod s

TABLE 9.2 Intraoperative EMG Activity vs. Postoperative Outcome in 213 Monitored Lumbar Spine Cases Postoperative symptoms

EMG activity

No EMG activity

Improved, 177

133 (75%)

44 (25%)

No change, 21

17 (81%)

4 (19%)

Worse, 14

14 (100%)

0 (0%)

From Gunnarson et al. (2).

Although there are many published anecdotal case reports touting the benefits of intraoperative EMG monitoring during lumbar spine surgery, there is a paucity of prospective outcome data. Some studies have shown reduced incidence of postoperative C5 radiculopathy after cervical decompression using EMG monitoring from the deltoid muscle, and cases detected during surgery with EMG monitoring were felt to have been milder, with a more rapid recovery because the surgeon was alerted immediately (3,4) (Table 9.4). These data can probably be extrapolated to nerve root monitoring during lumbar spine surgery. Monitored patients undergoing complex lumbosacral procedures for spinal cord tumors and tethered cord syndrome show abnormal EMG activity confined to the sphincter muscles in 33% cases (28% anal and 5% urethral sphincter), underscoring the importance of recording sphincter EMG in these cases (5).

Precautions Although EMG monitoring is sensitive for blunt mechanical irritation and injury to motor nerves, it may remain quiescent during sharp nerve transection. Furthermore, even after acute nerve transection the distal nerve stump can still be activated by mechanical irritation and electrical stimulation, and the presence of these ongoing evoked EMG responses may be mistakenly interpreted as evidence of nerve root continuity. Mechanical trauma is less likely to evoke neurotonic discharges from abnormal motor nerves, and EMG monitoring may miss additional intraoperative injury to nerve roots already partially injured by preexisting lumbosacral radiculopathy. Neurotonic discharges must be differentiated from other EMG findings that can occur during surgery without nerve injury to avoid false alarms. Spontaneous fibrillation potentials may be seen throughout surgery in patients with preexisting radiculopathy and denervation. They can be distinguished from neurotonic discharges because they occur continuously (even before critical phases of surgery) and fire regularly at slow rates (less than 15 Hz). Unlike neurotonic discharges, fibrillations are action potentials of single muscle fibers and not motor units, so they are simple triphasic or biphasic spikes (Figure 9.3A). Voluntary motor unit potentials can be seen under conditions of light anesthesia. These are semirhythmic motor unit potentials that usu-

TABLE 9.3 Sensitivity and Specificity of EMG vs. SEP Monitoring for Detecting 14 New Postoperative Deficits in 213 Monitored Lumbar Spine Cases

Sensitivity

Specificity

Positive predictive value

14/14 or 100%

23.7%

0.085

1.0

SEP monitoring (>50% change in amplitude) 4*/14 or 28.6%

94.7%

0.286

0.947

EMG monitoring (neurotonic discharges)

*All 4 patients with abnormal SEP also had abnormal EMG monitoring. From Gunnarson et al. (2).

Negative predictive value

Husain 09

1/17/08

12:00 PM

Page 143

CH A P T ER 9: Lumbos acral Surge r y TABLE 9.4 Incidence of Postoperative C5 Radiculopathy after Cervical Laminectomy for Cervical Myelopathy with and without Intraoperative EMG Monitoring from the Deltoid Muscle

Study

EMG monitoring

No EMG monitoring

Jimenez, 2005 (3)

1/116 (0.9%)

4/55 (7.3%)

Fan, 2002 (4)

2/68 (2.9%)

6/132 (4.5%)

ally occur simultaneously in muscles from multiple bilateral myotomes (Figure 9.3B). Electrocautery artifact is easy to recognize but

A

•

1 43

precludes concurrent EMG monitoring (Figure 9.3C). Needle electrode movement artifact can be mistaken for EMG activity when the surgeon is working close to the recording electrode (Figure 9.3D). Neurotonic discharges may occur without mechanical injury—for example, during cold-water irrigation within the surgical field. Finally, EMG cannot be monitored in the presence of total pharmacologic neuromuscular blockade because nerve activation from mechanical irritation must be transmitted across the neuromuscular junction in order to evoke identifiable myogenic responses.

C

B

D

FIGURE 9.3 Other EMG findings that can occur during surgery without nerve injury. A. Spontaneous fibrillation potentials may been seen throughout surgery in patients with preexisting radiculopathy and denervation. B. Voluntary motor unit potentials can be seen under conditions of light anesthesia. C. Electrocautery artifact. D. Needle electrode movement artifact. In each case the time base shown is 100ms/div and the display sensitivity is 50 µV/div..

Husain 09

144

1/17/08

•

12:00 PM

Page 144

SE C T I ON I I : C l i n i c a l Me thod s

STIMULUS-TRIGGERED EMG MONITORING TO VERIFY CORRECT PLACEMENT OF PEDICLE SCREWS Background Many spinal instrumentation systems use pedicle screws as a point of fixation. Holes are drilled blindly through the narrow pedicles into the vertebral body to facilitate pedicle screw placement. Cadaver studies have shown that as many as 20% of screws are misdirected, lying outside the bony pedicle wall (Figure 9.4). Clinical studies have detected postoperative symptoms from irritation or injury to the adjacent nerve roots by misplaced pedicle screws in 5% to 10% cases. Intraoperative fluoroscopy and radiographs have been used in an attempt to verify correct pedicle screw placement during surgery but are not always reliable, perhaps because of limited available projections. Computed tomography is more definitive, but of course is not available in real time in the operating room. Direct inspection and palpation of the medial wall of the instrumented pedicle by the surgeon is the “gold standard” test for misplaced screws during surgery, but doing this for every screw may necessitate unnecessary laminectomies and excessive operative times. Fortunately, stimulus-triggered EMG can be used to quickly verify correct pedicle screw placement in real time during surgery. Holes or screws that are correctly positioned within the pedicle wall are separated from the adjacent nerve roots by a cortical bony layer with high impedance to the passage of electrical current. (Figure 9.5B). However, a hole or screw that has perforated the bony pedicle wall will lie directly against adjacent nerve roots, and direct electrical stimulation of such misplaced holes and screws activates the adjacent nerve roots, evoking compound muscle action potential (CMAP) responses in muscles from the appropriate myotomes at lower stimulus intensities than those that lie completely contained within the pedicle wall (Figures 9.5A and 9.6A). Stimulus thresholds of less

FIGURE 9.4 Cadaver specimen showing a misplaced pedicle screw, a potential source of postoperative irritation or injury to the adjacent nerve root.

than 4 to 6 mA are suggestive of cortical bony perforation by pedicular instrumentation when the adjacent nerve root is healthy (6,7) (Table 9.5).

Technique Holes and screws are tested using a monopolar stimulating electrode with a remote anode needle electrode placed in the surgical wound and stimulus-triggered EMG recordings made from appropriate limb muscles. The stimulator is inserted directly into the pedicle hole after drilling. It is preferable to have the stimulator halfway into the hole, so that the uninsulated tip of the electrode is inside the pedicle close to the adjacent nerve root and not deep inside the vertebral body (Figure 9.5A). Screws are tested by touching the stimulator against the exposed shank or port of the screw, taking care to avoid any intervening soft tissue. It is important to test each pedicle hole and screw individually as each is drilled and inserted. A misplaced hole should be redirected and then retested before it is instrumented.

Husain 09

1/17/08

12:00 PM

Page 145

CH A P T ER 9: Lumbos acral Surge r y

•

1 45

FIGURE 9.5 Stimulus-triggered EMG for detecting pedicular wall breach. A monopolar stimulator is inserted into a pedicle hole or touched against a pedicle screw. A. Holes or screws that have perforated the bony pedicle wall will lie directly against adjacent nerve roots and stimulation activates the adjacent nerve root, evoking a CMAP response. B. Holes or screws that are correctly positioned within the pedicle wall are separated from the adjacent nerve roots by a cortical bony layer, with a high impedance to the passage of electrical current and no evoked CMAP responses.

A

B

FIGURE 9.6 CMAP responses from intraoperative stimulus-triggered EMG. A. CMAP response in tibialis anterior from stimulation of an L5 pedicle screw at 5.3 mA, suggesting pedicular breach. B. CMAP responses recorded after direct stimulation of exposed cauda equina (as a positive control) at 4 mA. The time base shown is 10 ms/div and the display sensitivity is 50 µV/div.

Husain 09

146

1/17/08

•

12:00 PM

Page 146

SE C T I ON I I : C l i n i c a l Me thod s

TABLE 9.5 Stimulus Thresholds for Normal Healthy Nerve Roots, Chronically Compressed Nerve Roots, and Normal and Misplaced Pedicle Holes and Screws*

Structure Normal nerve root Chronically compressed nerve root Normal hole

Stimulus threshold (mA) (0.2–5.7) 6.3–20 (16.5–44.3)

Normal screw

24 (12.1–35.9)

Misplaced hole

(1–6)

Misplaced screw

(1–6)

*Shown as mean and/or range. From Maguire et al. (6) and Holland et al. (9).

EMG using a handheld stimulator in the operative field can also be used to identify viable nerve roots during dissection of intraspinal dumbbell and foraminal tumors and spinal surgery for tethered cord release.

Outcome Data Based on a stimulus threshold of 6 mA or less, the sensitivity of stimulus-triggered EMG for identifying misplaced pedicle screws and hole is more than 90%, significantly exceeding the sensitivity of intraoperative radiography, which is only 63% (6). The probability of a pedicle screw breach increases with decreasing stimulus threshold (7) (Table 9.6).

Precautions Redirecting the hole will frequently close off the pedicular perforation with bone fragments pushed outward by the drill bit. Insertion of a pedicle screw may enlarge a previously normal hole causing a wall breach. Finally, some particular brands of pedicle screw have an unusually high electrical resistance based on their metallic composition, so just testing the screw (and not the preceding hole) may give rise to a false-negative result (8). To save time, each hole and screw can be initially tested at a searching stimulus intensity of 7 to 10 mA. If no CMAP response is identified at this intensity level, then the hole or screw is considered safe. If a CMAP response is identified at the initial searching intensity, the stimulating current is progressively reduced, using recurrent stimulation at 2 Hz, in order to determine the precise stimulus threshold required to evoke the response and hence the probability of pedicular perforation. Pedicle screws or holes with stimulus thresholds of less than 4 mA should be removed and/or redirected. Pedicle screws with stimulus thresholds between 4 and 6 mA are considered borderline and should probably be more closely inspected by the surgeon. The very same technique of stimulus-triggered

Because the expected result from a correctly placed pedicle screw or hole is negative (i.e., no response at the searching stimulus intensity), it’s always a good idea to test a positive control to be sure that the stimulator and amplifier are working properly. This can be accomplished by having the surgeon stimulate an exposed nerve root directly.(Figure 9.6B). It is important to recognize that the stimulus thresholds mentioned above refer to pedicular instrumentation at levels where the bone and adjacent nerve roots are normal and healthy. False-positive stimulus-triggered EMGs can be seen in patients with advanced osteoporosis, presumably because of thin cortical bone TABLE 9.6 Stimulus Threshold and Pedicle Screw Malposition (Confirmed by Palpation and/or Radiographs) for 4,587 Screws

Stimulus threshold

Probability of screw malposition

>8 mA

0.31%

4–8 mA

17.4%

50% ↓ FAR, ± ↑ WPS

17–22

Suppression

< 17

Severe Quantitative EEG Significant change

75% loss of total power

Propofol anesthesia

↑ relative delta power

Isoflurane anesthesia

SEF 90%

SEP Significant change

50% ↓ amplitude of cortical waveforms 5% ↑ latency of cortical waveforms SEP changes

CBF (mL/100 g/min)

Minor

< 50% ↓ amplitude of cortical waveforms

15–20

Major

> 50% ↓ amplitude of cortical waveforms, 5% ↑ latency of cortical waveforms

< 14

Severe

Loss of cortical waveforms, ↓ amplitude of subcortical waveforms

12–15

FAR = fast anterior dominant rhythmic activity; WPS = widespread persistant slow waves, CBF = cerebral blood flow; SEF = spectral edge frequency.

mL/100 g/min, significant loss of subcortical waveform amplitude may be seen. A major change should prompt the surgeon to insert a shunt. As with EEG, SEP changes resolve once the shunt is in place. Occasionally focal EEG and unilateral SEP changes can occur randomly during CEA—i.e., not related to CCC. Often these changes are transient and related to fluctuations in blood pressure or anesthetics. If they resolve spontaneously, they are not likely to be associated with postoperative morbidity. If presurgical focal EEG abnormalities are present, they may be accentuated with anesthesia, thus appearing to be new focal changes. These, too, are not associated with new postoperative deficits. Emboli may arise from the site of surgery and lodge in a distal vessel ipsilaterally. This can result in focal EEG and uni-

lateral SEP changes. These changes will be persistent; increasing the blood pressure may help resolve them. But if they persist, they will likely be associated with new postoperative deficits.

Sample CEA NIOM Case A 43-year-old man with an asymptomatic 80% left internal carotid artery stenosis and diffuse vascular disease underwent a left CEA with EEG and SEP NIOM. Upon initial clamping of the carotid artery, a marked loss of left more than right hemispheric QEEG total power was seen (Figure 16.11A). A Pruitt shunt was placed and all signals recovered (Figure 16.11B). The same pattern of signal loss and recovery was seen later in the procedure when the vessel was clamped again for

Husain 16

276

1/17/08

•

12:06 PM

Page 276

SE C T I ON I I : C l i n i c a l Me thod s

FIGURE 16.10 Example of a left-sided change (upper two channels) of QEEG during CCC.

shunt removal (Figure 16.11C). The amplitude of the right median nerve SEP cortical response was also dramatically reduced in association with the initial carotid artery clamping, as was the amplitude of the right tibial nerve SEP cortical response(Figure 16.12). All signals recovered and the patient remained neurologically intact after surgery (7).

Other Vascular Surgeries Surgeries involving distal branches of the internal carotid artery (anterior and middle cerebral and anterior and posterior communicating arteries) as well as arteries of the vertebrobasilar system require more complicated NIOM. This is not discussed in detail as monitoring for these surgeries is uncommonly per-

FIGURE 16.11 QEEG total power changes during a CEA case. Left more than right total power dropped upon clamping, followed by a recovery after shunting. At point “A,” the left carotid artery was clamped. A sudden drop in power was seen bilaterally, left worse than right. Point “B” marks institution of a shunt; notice prompt return of the waveforms to near normal levels. Toward the end of the case, at point “C,” the carotid artery is clamped again for shunt removal. There is sudden loss of total power, which returns after releasing the clamp.

formed. An overview of the modalities used most often is presented below. REEG, QEEG, and SEPs are used for most vascular surgeries, and brainstem auditory evoked potentials (BAEPs) are used for surgeries on the vertebrobasilar system. When vascular surgery involves the middle or posterior cerebral or posterior communicating arteries, contralateral tibial nerve SEPs are more sensitive than median nerve

Husain 16

1/17/08

12:06 PM

Page 277

CH A P T ER 16: Carotid Surge r y

•

2 77

FIGURE 16.12 Median nerve SEP changes during a right CEA. Right median nerve SEP response deteriorates after clamping with improvement after shunting without left median nerve SEP change.

SEPs. For surgeries on the A1 segment of the anterior cerebral and anterior communicating arteries, the contralateral tibial nerve SEP is most sensitive, followed by the ipsilateral tibial nerve SEPs, followed by the contralateral median nerve SEPs. Surgery on the A2 segment of the anterior cerebral artery is best monitored by contralateral tibial and median nerve SEPs. When the basilar or vertebral arteries are involved, both tibial and median nerve SEPs are useful, as are BAEPs. REEG and QEEG are used during surgery on all cerebral vessels.

Evidence Supporting Use of NIOM in Carotid Endarterectomy It is difficult to determine whether NIOM with EEG or SEP helps decrease the morbidity of CEA, as most surgeons shunt

the carotid artery if neurophysiologic changes are noted. However, a report from a center where none of the patients receive shunts regardless of EEG findings noted that 55 of 176 (31%) patients undergoing CEA had significant clamp-associated REEG changes (10). These changes resolved after the clamp was released in 36 (65%) patients. Five (9%) developed new neurologic deficits, and in two (1%) of these patients the deficits were permanent. Similarly, in a study of 312 CEA procedures, 28 (8%) had significant alterations of SEPs (13). In 24 of these patients, the abnormalities were completely reversed with shunting, and in another 2 patients the abnormalities were partially reversed. None of these patients had any postoperative deficits. In 2 patients in whom there was persistent absence of SEPs, contralateral hemiparesis was noted after surgery.

Husain 16

278

1/17/08

•

12:06 PM

Page 278

SE C T I ON I I : C l i n i c a l Me thod s

Another way to evaluate the efficacy of EEG and SEP monitoring is to evaluate their sensitivity and specificity in identifying cerebral ischemia. In a recent review, the sensitivity and specificity of REEG was noted to be 0.27 and 0.87. QEEG was slightly better, with sensitivity and specificity of 0.58 and 0.99. SEPs also seemed to be better than REEG, with a sensitivity and specificity of 0.52 and 0.98 (8). Multimodality monitoring may be more effective than any single modality alone.

for these modalities should also be made available. Details of obtaining BAEPs are discussed elsewhere in this book. The surgical team is responsible for informing the patient about the procedure and risks prior to the surgery. It is a good idea to include the remote risk of needle burn in the list of possible complications. If possible, the NIOM team should meet the patient prior to the administration of anesthesia and explain their roles in the procedure.

Procedure TECHNICAL CONSIDERATIONS Preparation In preparing for CEA or other surgery involving the carotid arteries, it is best to review the anatomy of the cerebral vasculature. A handy reference demonstrating the vascular anatomy around the circle of Willis is useful to have available. There are many patient-related conditions that may affect NIOM; it behooves the NIOM team to recognize these so that monitoring can be adequately adapted. Prior hemispheric stroke or demyelinating diseases may result in slowing of EEG frequencies at baseline. This must be kept in mind in interpreting focal slowing after CCC. The presence of peripheral neuropathy, neuromuscular disorders, myelopathy, and cerebral palsy may affect SEPs. Preoperative EEG and SEP studies done at least a day in advance of the surgery have the advantage of alerting the NIOIM team of preexisting abnormalities in neurophysiologic tests. An adequate number of electrodes should be available. The exact number and types of electrodes needed will depend on which modalities are monitored. Sterile, disposable, subdermal, stainless steel needle electrodes can be used for all standard procedures. They are optimal due to low impedances and easy application. If special techniques such as BAEPs are to be used, appropriate electrodes

Setup for a CEA or other carotid surgery will depend not only on the modalities to be monitored but also on how many channels are available on the NIOM machine. Older machines had 8 channels, greatly restricting the number of EEG and SEP channels that could be recorded. Now 32-channel machines are available; however, most machines used currently have 16 channels. At least four EEG channels are used; the most useful montage has been described earlier (F8 Cz, T4-Cz, C4-Cz, F4-Cz or F3-Cz, P4-Cz, C4-Cz, F7-Cz). A slightly different montage is used for aneurysm surgery (F1–C3, C3–O1, F2–C4, C4–O2). If extra channels are available or if separate machines are used for monitoring EEG and SEPs, a 16channel, anterior-posterior longitudinal montage should be used. For SEP monitoring, electrodes on the scalp (C3, Cz, C4), cervical spine (C5S), Erb’s point, and popliteal fossa are used. Stimulating electrodes for the median and tibial nerves are placed at the wrists and medial aspect of the ankles. The recording and stimulating electrodes for SEPs can be either of the needle or surface type. Anesthesia can affect both EEG and SEP NIOM. A discussion with the anesthesia team about the modalities being monitored and the best anesthetic regimen is important. Generally for CEA and other types of carotid surgeries, complete neuromuscular blockade is acceptable. Inhalation agents are also

Husain 16

1/17/08

12:06 PM

Page 279

CH A P T ER 16: Carotid Surge r y

acceptable in these cases. The NIOM team must keep in mind the effects of these agents on the EEG, as discussed earlier. If the minimum alveolar concentration (MAC) is increased much above 1, there may be suppression of the cortical waveforms of the SEPs. Propofol and barbiturates may also be used by the anesthesia team to induce burstsuppression (for cerebral protection). This makes interpretation of the EEG more difficult; however, it may be needed for cerebral protection. The NIOM team can help the anesthesia team by keeping them informed of the duration of the suppression; usually 7 to 10 seconds of suppression are required between “bursts” to protect the cortex. Propofol and barbiturates have minimal effect on the SEPs. Etomidate can induce burst suppression but causes a marked increase in amplitude of the EEG; this should be avoided because comparisons to baseline will no longer be valid. Whenever possible, it is best that a constant level of anesthesia be used, as bolus doses of any agent can produce sudden changes in both the EEG and SEPs, making interpretation difficult. Baseline NIOM data should always be obtained once the anesthetic regimen has been stabilized. It may be useful to print a sample of the EEG baseline prior to CCC for comparison to later tracings. Similarly, an SEP baseline should be obtained for later comparison. At least the cortical response (and preferably the subcortical and peripheral responses as well) should be displayed, as it is usually the highest in amplitude and therefore has the best signal-to-noise ratio. If a change in noted in the cortical responses, other responses will also have to be viewed for localization of the change. Most new NIOM machines allow display of multiple panels, so that REEG, QEEG, and SEPs can be viewed simultaneously. REEGs and QEEGs are acquired in a freerunning mode during the surgery. SEPs are obtained as frequently as possible, particularly during the riskiest segment of the procedures (CCC, coiling, embolization, etc).

•

2 79

Changes unrelated to surgical injury must be identified as such to avoid raising a false alarm. These changes can be induced by anesthesia, a drop in blood pressure, blood loss, peripheral nerve compression (positioning, blood pressure cuff), electrical noise, mechanical noise (drilling, suctioning), and physiologic artifacts [electrocardiogram (ECG), electromyogram (EMG), breathing]. It is the role of the NIOM team to recognize and troubleshoot the confounding factors in collaboration with the anesthesia team. Technical problems often occur in the operating room. The technologist, with the help of the neurophysiologist, should attempt to resolve these as quickly as possible. Problems with stimulation (for SEPs) may occur if the stimulating needles are pulled out during positioning. The stimulator itself may be malfunctioning. If the needles are pulled out, the NIOM machine will often alert the operator that an impedance limit has been reached. Changes in the SEPs will consist of signal loss, including loss of peripheral responses. Problems can also occur with the recording system. This results in excessive artifact seen in the EEG and very noisy traces or excessive input rejection of the amplifier for SEPs. The first step is to check impedances. If impedances are high, the corresponding electrodes should be checked to make sure they are making adequate contact. The connection of the electrode leads to the amplifier must also be checked. The head box connection holes must also be checked, as they may be defective; other holes can be tried. Finally, the electrode itself may need to be replaced, as the connection between the electrode and its lead may be broken. If impedances are acceptable, all previous problems are ruled out. The technologist must then concentrate on identifying and eliminating noise sources if possible. Devices that may be causing artifact and may be disconnected include the surgical bed, body warmers, blood warmers, and electrical islands. Prior to unplugging any device in the

Husain 16

280

1/17/08

•

12:06 PM

Page 280

SE C T I ON I I : C l i n i c a l Me thod s

operating room, the circulating nurse must be consulted to ensure that the device can be safely unplugged. The NIOM machine should not be plugged in islands or where multiple devices are plugged. If possible the EEG and SEP electrode wires should be braided, as this reduces artifact as well. Filtering specific electrical noise frequencies can also be attempted if other methods to reduce this artifact fail (i.e., 60 Hz filter). Excessive EMG contamination of the EEG or SEPs can be reduced by asking the anesthesia team to administer neuromuscular blocking agents. When the NIOM technologist sees certain changes, they should be discussed with the neurophysiologist and anesthesia team before raising an alarm for the surgeon. This includes any global change affecting either EEG or SEPs and is not correlated with critical steps of the procedure. It must be determined as soon as possible if there have been changes in anesthetic parameters, blood pressure, temperature, etc. If the cause of these changes cannot be attributed to a nonsurgical cause or one of the alarm criteria described earlier is met, the surgeon must be notified immediately. Communication between the NIOM, surgery, and anesthesia teams is critical for optimal monitoring. Surgeons should announce every step of the surgery or any ongoing adverse event. The anesthesia team must inform of any parameter change or fluctuation in vital signs. The NIOM team must immediately inform the surgeon about any change that may imply ischemia or injury or discuss with the anesthesia team any nonsurgical change. Documentation of all events is extremely important.

Postprocedure Needle electrodes must be removed cautiously, avoiding injury to both the patient and others working in the operating room. The needle electrodes must be discarded in an appropriate container. In case of accidental

needle-stick injury, hospital protocols for such injuries should be followed. All tape fragments must be removed carefully as well. Surface electrodes, if reuseable, should be appropriately cleaned and disinfected for future use. Burn injuries are very rare, but when they occur, they should be reported and documented. Equipment should undergo scheduled maintenance by the biomedical department of the hospital. Technical malfunctions should be reported immediately and will require an immediate assessment by the engineers. After surgery, a report of the NIOM procedure should be written providing relevant data and sequence of events for accuracy. This report is also of important medicolegal value: whatever is not reported is considered not to have been done. Lack of thorough reporting may be interpreted as “sloppy” and unprofessional monitoring. Documentation must include patient demographics, diagnosis, relevant clinical or complementary data, and any adverse event or change in NIOM. Consider including baseline and closing data printouts, including SEP curve stacks and numerical data, as part of the patient’s NIOM chart. This data can also be stored digitally, depending on the NIOM machine’s capability.

CONCLUSIONS NIOM during CEA and other types of carotid surgery can provide the surgeon with critical information that may modify the surgical procedure. With assistance from the NIOM team, the surgeon may elect to shunt the carotid artery during CEA or may modify the clip placed on an aneurysm. REEG and QEEG analysis during these surgeries is critical for assessment of cortical perfusion. With the addition of SEP monitoring, subcortical ischemia can also be detected. Recognizing appropriate alarm criteria is important, as instantaneous feedback to the surgeon is mandatory. A complementary working environment between the

Husain 16

1/17/08

12:06 PM

Page 281

CH A P T ER 16: Carotid Surge r y

NIOM, surgery, and anesthesia teams will ensure the best monitoring. 9.

REFERENCES 1. Netter FH. Nervous System: Neurologic and Neuromuscular Disorders. West Caldwell, NJ: Ciba-Geigy, 1992. 2. Netter FH. Atlas of Human Anatomy, 3rd ed. Teterboro, NJ: Icon Learning Systems, 2003. 3. Yadav JS, Wholey MH, Kuntz RE, et al. Protected carotid-artery stenting versus endarterectomy in high-risk patients. N Engl J Med 2004;351:1493–1501. 4. Allain R, Marone LK, Meltzer J, Jeyabalan G. Carotid endarterectomy. Int Anesthesiol Clin 2005;43:15–38. 5. Liu AY, Lopez JR, Do HM, et al. Neurophysiological monitoring in the endovascular therapy of aneurysms. AJNR Am J Neuroradiol 2003;24:1520–1527. 6. Babikian VL, Cantelmo NL. Cerebrovascular monitoring during carotid endarterectomy. Stroke 2000;31:1799–1801. 7. Minahan RE. Intraoperative monitoring. Neurologist 2002;8:209–226. 8. Florence G, Guerit JM, Gueguen B. Electroencephalography (EEG) and somatosensory

10.

11.

12.

13.

•

2 81

evoked potentials (SEP) to prevent cerebral ischaemia in the operating room. Neurophysiol Clin 2004;34:17–32. Laman DM, van der Reijden CS, Wieneke GH, et al. EEG evidence for shunt requirement during carotid endarterectomy: optimal EEG derivations with respect to frequency bands and anesthetic regimen. J Clin Neurophysiol 2001;18:353–363. Blume WT, Sharbrough FW. EEG monitoring during carotid endarterectomy and open heart surgery. In: Niedermeyer E, Da Silva FL, eds. Electroencephalography: Basic Principles, Clinical Applications, and Related Fields. Baltimore: Williams & Wilkins, 1999:797–805. Sloan TB. Anesthetic effects on electrophysiologic recordings. J Clin Neurophysiol 1998; 15:217–226. Laman DM, Wieneke GH, van Duijn H, et al. QEEG changes during carotid clamping in carotid endarterectomy: spectral edge frequency parameters and relative band power parameters. J Clin Neurophysiol 2005;22: 244–252. Amantini A, Bartelli M, de Scisciolo G, et al. Monitoring of somatosensory evoked potentials during carotid endarterectomy. J Neurol 1992;239:241–247.

Husain 16

1/17/08

12:06 PM

Page 282

Husain 17

1/17/08

12:06 PM

17

Page 283

Epilepsy Surgery

William O. Tatum, IV Fernando L. Vale Kumar U. Anthony

M

Once the site of ictal onset has been identified, neurophysiologic intraoperative monitoring (NIOM) is often necessary during the resective surgery. NIOM in these cases includes electrocorticography (ECOG) and functional brain mapping. ECOG is a neurophysiologic technique that records cortical electrical potentials directly from the surface of the brain. It was pioneered during brief exposures of the brain in the operating room in an effort to determine the site and borders of epileptogenicity and offer predictive information about postsurgical outcome. Functional brain mapping using electrical stimulation is helpful in defining and confining the excision to noneloquent cortex. These two techniques are further discussed in this chapter.

ore than 3 million people in North America have epilepsy, and 20% to 40% have seizures that remain uncontrolled with antiepileptic drugs (AEDs). Most adults with seizures refractory to AEDs have localization-related epilepsy (LRE) and recurrent complex partial seizures. Up to 50% of these patients may be candidates for neurosurgical intervention. Surgical candidates must have a definitive diagnosis of LRE, fail at least two or three adequate trials of an appropriate AED, be impaired by their seizures, and be motivated to undergo epilepsy surgery. The role of the neurophysiologist in epilepsy surgery is to localize the epileptogenic zone, elucidate the functional anatomy of the neurophysiologic generator, and predict the outcome after epilepsy surgery. Localization of the epileptogenic zone is based upon concordance of the preoperative evaluation, which includes interictal and ictal electroencephalograms (EEGs) recording from the scalp and various types of neuroimaging. However, scalp EEG localization may be difficult when the epileptogenic zone is in a deep-seated location, small in size, associated with rapid low-amplitude rhythmic ictal discharges, or limited by movement artifact that obscures the recording. In these situations, invasive recordings may be necessary to clarify the site of ictal origin.

ANATOMY AND PHYSIOLOGY The brain is compartmentalized into the frontal, parietal, occipital, and temporal lobes. Much of the cerebral cortex has surface area that is “buried” beneath the skull and is therefore inaccessible to scalp EEG recording. Additionally, individual variations in skull thickness create differences in scalp EEG potentials from patient to patient and between the left and right hemicranium. Brain 283

Husain 17

284

1/17/08

•

12:06 PM

Page 284

SE C T I ON I I : C l i n i c a l Me thod s

functional compartmentalization also varies from patient to patient. This is especially important in defining a cortical map of functional neuroanatomy. Cortical representation of a body region is illustrated best by an individualized homunculus. The homunculus is a functional, three-dimensional, disproportionate human caricature with regional accentuations (i.e., face and hand enlargements) superimposed on the surface of the brain. In the homunculus, the head and hands are positioned inferolaterally and the legs superomedially upon the midsurface of the brain. The amount of cortical representation depends upon individualized development or previous injury restricting regions of involvement. The central sulcus is a critical landmark that helps to identify the principal sites of motor and sensory function. The primary motor cortex is demarcated by the precentral gyrus residing at the posterior margin of the frontal lobe. The primary somatosensory cortex is represented by the postcentral gyrus located in the anteriormost parietal lobe; it integrates sensory functions, although it also has representation from the corticospinal tracts to impart motor function as well. The location of language function is dependent on cerebral dominance. The principal functions of the temporal lobe are the functional integration of language within the dominant hemisphere and mediating functions responsible for memory and learning. Language function is processed by reception of written or spoken words through cortex located at the supramarginal and angular gyrus of the anterior parietal lobe. Expression of language is effected through the Broca area in the posterior-inferior frontal lobe. The more posterior temporal-occipital lobe connections and occipital lobe serve visual functions mediated by the primary visual cortex. The temporal lobe is frequently the site of epilepsy onset (temporal lobe epilepsy, or TLE); consequently it is a common site for surgical intervention. The borders of the temporal lobe are the Sylvian fissure superiorly,

the middle cranial fossa inferiorly, and the sphenoid bone anteriorly, with an approximate border from the parietal and occipital lobes posteriorly. The hippocampal formation is oriented in the inferomesial portion of the temporal lobe along its anteroposterior extent and consists of the parahippocampal gyrus, subiculum, hippocampal sulcus, dentate gyrus, alveus, and fimbriae, creating convolutions of the brain that mimic the appearance of a seahorse. Scalp EEG measurement of the epileptogenic zone represents a two-dimensional projection from a three-dimensional source. A single interictal epileptiform discharge (IED) detected at the scalp electrodes reflects large synchronous electrical discharges of several million neurons. Therefore the cortical potentials produced at the scalp are often not representative of small interictal or ictal sources. Furthermore, dipole localization using scalp EEG to infer the origins of an IED creates the inverse problem of source localization due to volume conduction inherent in scalp EEGs, which “scatters” the electrical activity from the intrinsic sources, preventing accurate source prediction. In addition, most of the human cortex is buried deep beneath the scalp surface and is far removed from the detection of scalp EEGs (Figure 17.1). Hence, intracranial EEGs may clarify neurophysiologic localization when noninvasive investigations are discordant or nonlocalizing on scalp EEGs (Figure 17.2).

PATHOLOGY Resection of the epileptogenic zone that represents the anatomic site of epileptogenesis is essential to a favorable surgical outcome for patients with intractable seizures. While the pathology of TLE is most often hippocampal sclerosis, the pathology of extratemporal epilepsy is much more varied. Slow-growing tumors such as low-grade astrocytomas, oligodendrogliomas, gangliogliomas, and

Husain 17

1/17/08

12:06 PM

Page 285

CH A P T ER 17: Epile ps y Surge r y

•

2 85

FIGURE 17.1 Concomitant scalp and intracranial EEG demonstrating repetitive right temporal IEDs, which are seen on the intracranial channels but not on the scalp channels in a patient with temporal lobe epilepsy. LAT = left anterior temporal; LMT = left midtemporal; RAT = right anterior temporal; RMT = right midtemporal; lower channels = scalp EEG. [Reproduced with permission from Tatum WO, Benbadis SR (14).]

FIGURE 17.2 Auras that were not detected on scalp EEG were recorded from the wall of an area of encephalomalacia in a patient with symptomatic extratemporal epilepsy. An array of two right temporal strips, a strip within the cyst wall, and a grid were implanted unilaterally. G = grid; CW = cyst wall; RAT = right anterior temporal; RPT = right posterior temporal.

Husain 17

1/17/08

286

•

12:06 PM

Page 286

SE C T I ON I I : C l i n i c a l Me thod s

dysembryoplastic neuroepithelial tumors are frequently discovered. Malformations of the brain due to abnormal cortical development are common pathologic substrates and are being visualized with increasing frequency owing to advancements in neuroimaging. Cortical dysplasia represents a wide range of neuronal migration disorders that create a structural and neurophysiologic disturbance. Vascular malformations such as cavernous vascular malformations and arteriovenous malformations may lead to intractable epilepsy. Encephalomalacia of various etiologies—including trauma, infection, and cerebral infarction in addition to other causes—may also produce seizures.

ECOG Background Hans Berger initially performed ECOG in the late 1920s, when he recorded human EEGs through craniotomy sites with intracortical needle electrodes (1). In the mid-1930s, EEG was introduced into the operating room (2). ECOG is best appreciated as a technique that is used in the operating room to record EEGs from surgically exposed brain. ECOG is composed of the same combination of cortical rhythms found on scalp recording. However, the amplitudes of ECOG are 10-fold greater than at the level of the scalp. Faster frequencies enhanced by medications may appear similar to pathologic polyspike discharges when compared with the scalp EEG owing to the absence of the skull and scalp, which normally attenuates the voltage of the low-amplitude fast frequencies. Regional differences in cortical representation of the different frequencies are more discrete and therefore more apparent with ECOG. In addition, normal variations and variants, such as the mu rhythm, are more precisely demarcated and restricted on ECOG, mimicking abnormal IEDs.

Intraoperative Recording Intraoperative ECOG refers to the recording of neurophysiologic potentials directly from the brain within the confines of the operating room. When used as a presurgical tool, intraoperative ECOG is limited because of the brief opportunity for recording during surgery, and preoperative localization information creates a bias toward exposure. However, there are benefits to intraoperative ECOG, with the potential to avert risks from the initial surgical procedure, move and direct electrode locations within the surgical field, and perform intraoperative ECOG with cortical stimulation for mapping. ECOG is recorded with standard EEG equipment brought to the operating room. Electrodes of various types and styles may be used. Seizures are seldom recorded in utilizing ECOG; the IED provide the bulk of the information obtained within the operating room. Intraoperative ECOG has been used in an effort to localize the site of epileptogenicity through the demonstration of IED persistence, frequency, and distribution. Because neural tissue is complex, pitfalls in utilizing IED as a marker for seizure onset lie in the fact that the functional relationships between IED and seizure onset are not always commensurate. Additionally, the epileptogenic zone and the sites of IED formation may be of dissimilar sizes; widespread discharges may be noted despite a smaller zone of epileptogenesis. Furthermore, spikes and sharp waves that are propagated or even volumeconducted are not readily distinguishable from spikes and sharp waves involved in the primary generator. Hence, seizures may arise from sites outside the ECOG sites of interictal epileptiform abnormalities.

Extraoperative Recording Interest in intraoperative ECOG in guiding resection of dysfunctional cortex has now evolved to include advances made within neuroradiology as well as digital instruments

Husain 17

1/17/08

12:06 PM

Page 287

CH A P T ER 17: Epile ps y Surge r y

available for long-term video–EEG monitoring. With the need for invasive recordings of seizures and functional mapping of eloquent cortex, extraoperative ECOG has become the foundation for excisional epilepsy surgery when intracranial electrodes are necessary. While anesthesia is not utilized outside the operating room, analgesia is typically required. Extraoperative ECOG allows clarification of briefer intraoperative recordings by increasing the length of recording and obtaining information from intracranial ictal EEGs. It also allows more prolonged cortical mapping time when necessary. Intraparenchymal (intracortical or depth) or extraparenchymal (subdural strip/grid) recordings are used during the intracranial phase of epilepsy monitoring. Deep-seated anatomic structures—including the hippocampus, amygdala, and subcortical abnormalities within the brain—may serve as generators for seizures. As such, volume conduction of these electrical fields from deep generators may interfere with precise localization at the level of the scalp. New frameless systems for the placement of depth electrodes near these generators allow for more accurate localization.

Electrodes The initial intraoperative ECOG pioneered in Montreal, Canada, by Jasper and Penfield used a series of eight insulated silver wires with chloride ball tips as recording electrodes. They were capable of being redirected to different sites of the exposed brain. Artifactcontaminated reference recordings quickly gave rise to bipolar montages. Today, subdural electrodes constitute the main method of recording ECOG and are available from several electrode-dedicated vendors. Subdural electrodes are available in various configurations that comprise grid and strip electrodes involving multiple contacts and sizes that can be tailored for individual implantation (Figures 17.3 and 17.4). Stainless steel and

•

2 87

FIGURE 17.3 Photograph of an 8 by 8 contact subdural grid electrode.

platinum alloys are the principal compounds composing the electrodes; they are embedded in a clear malleable silicone or polyurethane array to permit visualization of underlying cortical structures during placement (Figure 17.5). A variety of electrode styles is available; these typically include strips with 2 to 8 contacts and grids that contain 4 to 64 contacts. Each contact may vary in diameter but commonly is 5 mm wide, and they are spaced 1 cm apart. Individual sizes and spacing depend upon clinical requirement as well as manufacturing capabilities. Subdural electrodes are placed over the site of cortex that is potentially epileptogenic or eloquent for an essential function (i.e., primary

FIGURE 17.4 Photograph of a six-contact subdural strip electrode. Each contact is separated by 1 cm.

Husain 17

288

1/17/08

•

12:06 PM

Page 288

SE C T I ON I I : C l i n i c a l Me thod s

FIGURE 17.6 Photograph of a depth electrode. This electrode has eight contacts.

FIGURE 17.5 Operative placement of a 4 by 5 contact subdural grid in a patient with medically intractable epilepsy. Note the retracted dura mater at the top and the cables connecting to the EEG at the bottom.

sensory, motor, visual, or language cortex). The advantages of subdural electrodes over depth electrodes are that they are less likely to create morbidity, are easier to implant, and can cover a larger cortical surface. Depth electrodes are intraparenchymal electrodes that sample a small surrounding area of brain at the site of placement (Figure 17.6). This form of stereo–EEG recording provides a limited sampling and represents the most invasive of all the commonly used electrodes for direct recording from the brain. Depth electrodes have the advantage of allowing more precise localization from the mesial temporal structures when seizures arise from the amygdala and hippocampus. Other less invasive electrodes include foramen ovale and epidural electrodes for localizing the epileptogenic zone to one mesial temporal region.

Technique After the patient is taken to the operating room and has undergone general anesthesia, the head is secured using Mayfield pin holders, shaved, prepped, and draped in a sterile fashion. The site or sites of operation, positioning, and incisions are individualized for placement of electrodes, cortical stimulation,

or resection. A craniotomy (or craniectomies) is preformed to expose the dura mater, which is then opened to expose the cortical surface for placement of electrodes or for intraoperative cortical stimulation. The placement of subdural grid electrodes requires a craniotomy, while strip electrodes may be placed through individual burr holes. After the neurophysiologic evaluation is completed, resection of brain tissue is performed using deep retractors, and adequate visualization is obtained using the operating microscope. Corticectomy and resection of the targeted regions are then performed. Tissue cultures are taken from the surgical wound site and antibiotics used to reduce the risk of unexpected infection. Thereafter, Surgicel is applied to the brain surface, retractors and the operating microscope are removed, the dura mater is reapproximated with sutures, and the bone flap is secured in place. Surrounding soft tissues are also reapproximated with sutures, and the skin is stapled. Perioperative antibiotics and steroids are administered and the patient is released to the anesthesia service for recovery. Recording ECOG inside and outside the operating room uses the same technique. A preoperative plan is developed prior to surgery. Selection of electrode type(s) to be used and plans for an implanted electrode array are required. Electrode placement occurs under aseptic conditions. The leads from the individual electrodes connect to the EEG machine for direct recording in the operating room and reconnected to cable (or radio) telemetry for further extraoperative recording. For extraoperative recording, the electrode leads are tunneled

Husain 17

1/17/08

12:06 PM

Page 289

CH A P T ER 17: Epile ps y Surge r y

under the skin and exit through a stab wound a short distance away from the craniotomy site. This helps minimize the risk of cerebrospinal fluid leakage and infection. Following completion of epilepsy monitoring with or without extraoperative functional mapping, the electrodes are removed during a second surgery. Sterile technique during placement and throughout the time of ECOG is essential. The risk of infection and hemorrhage with insertion of subdural strip electrodes is less than 1%, while subdural grid electrodes carry a higher risk of complications. Besides infection, transient neurologic deficits, hematoma, increased intracranial pressure, cerebral infarction, and herniation can occur due to the mass effect, which may be created by placement of the electrodes. The greater the number of electrodes implanted, the longer the recording time, the older the patient, the greater the number of skull breaches, or dominant hemispheric implantation, the greater the risk of complications. Unlike subdural electrodes, depth electrodes penetrate brain parenchyma and may cause intracerebral hemorrhage when placed near vascular structures. Examination of resected tissue has demonstrated gliosis and cystic degeneration around the trajectory of the depth electrode. Typically, however, the risk of intracranial bleeding or infection is only 0.5% to 4% (3). In the localization of seizure foci, intraoperative surgical navigation systems have been utilized to coregister subdural electrodes to regions of known radiographic pathology. This provides a link between noninvasive and invasive data to further develop the relationship between brain function and anatomic variability.

Interpretation Epileptiform Abnormalities Spikes are more evident on ECOG than the scalp EEG (Figure 17.1). About 25% of patients with TLE have IEDs on ECOG that

•

2 89

are not noted on the scalp EEG (2). Additionally, the area containing IEDs with ECOG is frequently larger than that obtained with scalp recordings. Compared to scalprecorded IEDs, ECOG-recorded IEDs have greater amplitude (500 to 1000 µV) and are shorter in duration, sometimes even less than 20 ms. Furthermore, ECOG IEDs may appear in several regions distant to the site of epileptogenicity. When patients have few or rare IEDs on ECOG, activation techniques are used to augment interictal abnormalities. Unlike scalp recordings that use hyperventilation, photic stimulation, and sleep as activating techniques, pharmacologic activation is used during ECOG. Methohexital and thiopental are used most frequently. Abnormalities that can be induced include focal reduction of fast frequencies, induction of focal slow activity, and an increase in the number, extent, and frequency of IEDs. However, IEDs noted with activation may extend beyond the proposed site of resection and region of neuroanatomic abnormality. Consequently the clinical usefulness of drug-induced IED in guiding resective surgery is limited. Information obtained from ECOG has been used to tailor resections of epileptogenic tissue. However, resection of the entire region with IEDs is not essential to render patients seizure-free. This is because the IEDs arise not only from the epileptogenic zone (which is resected) but also from adjacent regions of secondary cortical neurophysiologic involvement (which do not need to be resected). Therefore some investigators have found no additional information using ECOG beyond that obtained from the preoperative scalp video–EEG monitoring (3,4). On the other hand, areas of high density computer-detected spike discharges on ECOG have been suspected to be areas in the primary epileptogenic zone, with resection resulting in a more favorable surgical outcome (5). Dysplastic cortex often produces a unique pattern of epileptiform abnormalities on

Husain 17

290

1/17/08

•

12:06 PM

Page 290

SE C T I ON I I : C l i n i c a l Me thod s

ECOG. Prominent, widespread IEDs that may be very complex, repetitive, and have polyspike morphology are typically seen. Nearly continuous spikes and periodic spikes can also occur in long runs. It has been suggested that beyond localizing dysfunctional cortex, the morphology and topography of the IEDs may be predictive of pathology in cortical dysplasia.

Nonepileptiform Abnormalities Nonepileptiform abnormalities identified by ECOG include abnormalities of a variety of frequencies. A reduction of certain frequencies (i.e., beta activity) during spontaneous or pharmacologic activation identifies dysfunctional cortex. However, regions of low amplitudes on ECOG may reflect the technical limitations from bipolar recording of similar generators. Additionally, low-amplitude activity may signify proximity to vascular or noncerebral materials (Gelfoam, pledglets, saline irrigation, etc.) rather than pathology. Focal slow activity may be recorded on ECOG. This activity usually reflects dysfunctional cortex and may be helpful when IEDs are not present on the preexcision ECOG, although its lack of specificity limits its usefulness. With spectral analysis of ECOG in 40 patients, the area of maximal delta slowing coincided with the site of maximal spike activity in approximately 50% of patients in one study (5). However, given the difficulty in separating the surgical effects from underlying pathology after the start of operation, interpretation of intraoperative focal slowing merits suspicion of a nonepileptogenic source. Event related desynchronization of the normal alpha and beta frequencies has been used to detect differences during functional activation of sensorimotor cortex. Broad somatotopic networks demonstrate functional overlap of different body parts for alpha and beta frequencies. Comparatively, the topographic patterns of gamma activity (30 to 100 Hz) in the lower or upper limits of the bandwidth appear more discrete with respect to somatotopic specificity, appearing over the contralat-

eral somatosensory cortex during limb movement. In addition, coherence patterns identified with ECOG have suggested characteristic patterns that may help define the anatomy of individual brain regions. In defining the central sulcus, low phase coherence may decrease, while high phase-shift coherences increase. Importantly, pathologic alterations of brain function may be suggested by certain abnormalities. High-frequency oscillations ranging from 100 to 250 Hz (ripples) or 250 to 500 Hz (fast ripples) have been recorded using intracranial electrodes. Electrographic seizure onset has been associated with ripples; thus they have localizing value. Fast ripples occur preferentially in the hippocampus ipsilateral to the seizure onset, and their generators may be important in identification and mapping of the epileptogenic process.

Evidence of Usefulness Preexcision ECOG The evidence to suggest that the preexcision ECOG helps to determine the degree of resection for temporal, extratemporal, lesional, and nonlesional surgeries to provide a favorable clinical outcome has been limited (5,6). Early reports in adults and children with ECOG-guided resection of slow-growing glial lesions noted more favorable results with younger patients (7). A small series of patients with symptomatic intractable epilepsy caused by gangliogliomas reported improved outcome when ECOG-guided resection was performed compared with those who simply underwent a lesionectomy (5). ECOG often reveals widespread abnormalities beyond the site seen on neuroimaging. However, cortical dysplasias and glioneuronal tumors often cause localized or regional continuous spiking, bursts, and recruiting discharges (8). ECOG often reveals widespread abnormalities beyond the site seen on neuroimaging. A complete resection of the epileptiform abnormalities correlates with a favorable surgical outcome in some series (8).

Husain 17

1/17/08

12:06 PM

Page 291

CH A P T ER 17: Epile ps y Surge r y

Other investigators, however, report that more than one independent spike focus noted on preexcision ECOG predicts a poor postoperative outcome. Studies of nonlesional TLEs have not been able to identify favorable predictors for successful postoperative outcome (3–6). The correlation between underlying pathology and the location of IEDs on ECOG has not been reliable (3). This is likely due to frontal, parasagittal, and occipital foci having a more widespread neuronal network, leading to the propagation of IEDs widely over the hemisphere (or even the contralateral hemisphere).

Postexcision ECOG As with the preexcision ECOG, evidence for a beneficial role of the postexcision ECOG to predict surgical outcome has been inconsistent. Early reports suggested that the postexcision ECOG had predictive outcome potential when preexcision IEDs were no longer present after resection. One early study with 5-year follow-up demonstrated a significant difference when the postexcision ECOG did not have IEDs. When postexcision IEDs were present, only 36% of 104 patients had a good outcome (5). Other studies have demonstrated that residual postexcision spikes have no predictive value. This may be especially true of hippocampal or deep-seated lesions. One case series of pure lesionectomy (normal tissue at the edge of resection) patients reported seizure freedom to be independent of spike distribution or even spike presence on ECOG before or after resection (9). Additionally, another study found no difference with the use of ECOG in patients with TLE, though resection margins may have encompassed the majority of the IEDs (5). Some locations, such as the posterior parahippocampal gyrus and insular cortex, appear to lack prognostic significance when persistent spiking is encountered on the postexcision ECOG (2). IEDs may appear on the ECOG following initial resection when none were present prior to resection. Postexcision activation of spikes

•

2 91

is more benign than the presence of residual spikes unaltered by the resection and may indicate injury potentials from surgical manipulation. Discharges that remain unaltered after resection may carry a poorer prognosis for a seizure-free outcome, especially in frontal lobe epilepsy when three or more gyri contain IEDs. However, in a recent study, 80 patients with mesial TLE underwent a resection tailored to remove up to 7 cm of tissue, depending upon the presence of neocortical spikes and eloquent cortex. Interestingly, the presence of neocortical spikes was found to be associated with a more favorable prognosis for postoperative cognitive outcome (10).

FUNCTIONAL BRAIN MAPPING Electrical Stimulation Electrical stimulation of the brain has been primarily used for diagnostic purposes, although therapeutic uses also exist. The latter are beyond the scope of this chapter. Intracranial electrodes that are surgically implanted in the form of either strips or grids are capable of performing not only ECOG but also electrical stimulation of the cortex for functional brain mapping. Electrical stimulation can be useful for providing an individualized functional map that identifies areas of eloquent cortex. However, if a site on the map involved with a particular function is accidentally sacrificed during surgery, it does not necessarily imply that there will be a permanent deficit of that function. Often other neural networks are available to compensate for the lost function. Young children are able to demonstrate such plasticity better than adults. Stimulation of brain structures has been performed by direct application of electrical current to both cortical and subcortical structures.

Utility Electrical stimulation has been performed in every lobe of the brain. It may be performed

Husain 17

292

1/17/08

•

12:06 PM

Page 292

SE C T I ON I I : C l i n i c a l Me thod s

extraoperatively at the bedside through implanted electrode arrays or intraoperatively using handheld probes directed to specific cortical areas of interest by the neurosurgeon. Low-frequency stimulation (i.e., 1 Hz) is less likely to induce seizures than high-frequency (50 Hz) stimulation. However, even with lowfrequency stimulation, seizure induction has been noted with stimulation of temporal lobe white matter as well as the neocortical gray matter (11). Low-frequency electrical stimulation of the temporal lobe with simultaneous ECOG results in a gradual increase in spikes with or without the occurrence of low-voltage fast activity. This has been reported to help define the epileptogenic zone (11). Electrical stimulation to define the cortical representation of sensorimotor or language function of the dominant hemisphere has been a principal goal of brain mapping. Individual variations of the classic human homunculus are common, especially when lesions are present. One study found variation in the organization of primary motor cortex in 19.4% of 36 patients (12). In addition, functional overlap between two different areas was found in 11.1% of patients. The responses obtained from electrical stimulation of eloquent and “silent” regions of the brain allow the neurophysiologist to design a pictorial map of individualized cortical function that is task-specific. Thus this type of mapping is critical for optimizing postoperative outcomes.

Localization for Reproduction of Symptoms and Signs Various symptoms and signs can be reproduced by electrical stimulation. Negative and positive clinical correlates may be noted during functional brain mapping through electrical stimulation of the brain. Stimulation of negative motor areas within the frontal lobe, such as the supplementary motor area, may cause interruption fine motor movements, focal negative myoclonus, or bilateral atonia. Interruption of tongue, finger, or toe movement may be induced without loss of muscle

tone when the patient is asked to perform continuous movement tasks during testing. Numbness or scotomata can occur with stimulation of somatosensory and visual cortices. Stimulation of language areas may result in naming difficulties (as noted during the picture identification task), speech arrest, or difficulties in comprehension during reading. Positive motor phenomena include tonic or clonic contraction of a group of muscles and may be subjectively detected through an increase in tone or objectively noted on direct visualization. Tingling or phosphenes reflect positive involvement of the somatosensory or visual systems. In the low postcentral region, cortical landmarks determined by localizing tongue and face sensation are more reliable than stereotactic coordinates of the same area. Experiential or psychic symptomatology, visual imagery, and memory recall have been demonstrated during electrical stimulation of the mesial temporal lobe. If negative or positive phenomena are produced with stimulation, a second trial should be conducted for validation. This allows development of a functional map of the cortex. Additionally, by reproducing symptoms characteristic of the patient’s spontaneous events, the site of seizure onset can be better localized. Beyond defining the clinical function of neural tissue, electrical stimulation using depth electrodes for the purpose of triggering partial seizures has been performed. Identifying the epileptogenic zone from the irritative zone is done by reproducing the entire habitual seizure semiology. This technique has been inconsistent for specifically reproducing the habitual seizure. Additionally, more than one site from the ipsilateral and contralateral hemispheres may be capable of reproducing the same aura (11). Therefore electrical reproduction of the habitual aura lacks localizing and lateralizing specificity. However, when a typical aura is elicited by brain stimulation, the lobe of origin of the seizure can be established with some accuracy.

Husain 17

1/17/08

12:06 PM

Page 293

CH A P T ER 17: Epile ps y Surge r y

Localization of Afterdischarges Afterdischarges (ADs) are electrical events that represent elicited epileptiform activity similar to a restricted focal electrographic seizure. ECOG is used in conjunction with cortical stimulation to elicit and detect ADs. ADs have a threshold that reflects the degree of ease of seizure generation. They may arise from normal and pathologic brain tissue. Although the morphology remains stable in a given region of the brain, there is prominent variability from region to region and from patient to patient (12). The ADs may appear either as repetitive IEDs or brief rhythmic discharges. They serve as an electrographic warning sign that a seizure may be impending. No specific morphology predicts an underlying pathology, and a wide range of patterns with different frequencies may occur. If symptoms are produced with the ADs, it does not confirm proximity to functional cortex, as symptoms may be produced from propagated electrical discharges (11,12). Symptoms produced with and without ADs spreading beyond the site of stimulation have included not only sensory and motor responses but also visceral sensory, autonomic, thermoregulatory, experiential, and vocalization symptoms. Furthermore, specific clinical responses can often be elicited from more than one site, frequently from noncontiguous areas in the same or both hemispheres (11). AD thresholds reflect the degree of electrical current that is required to generate an AD. Thresholds do not appear to be AED concentration-dependent. The ADs vary in morphology, duration, and location and demonstrate greater excitability and lower thresholds at the site of seizure onset. Thresholds evoking ADs that are lower or more prolonged may carry a greater likelihood of corresponding to the site of spontaneous seizure onset, while morphology is probably less predictive (2). Thresholds will vary from patient to patient, and even the same site in the same patient may vary from day to day. Furthermore, AD

•

2 93

thresholds may tend to increase with subsequent stimulations or following prolonged ADs (2). Thresholds also depend on the location of the brain that is stimulated, with the mesial temporal lobe typically having the lowest threshold (13). Extratemporal neocortex has a higher threshold, and the motor cortex has the highest. Depth electrodes are amenable to cortical stimulation and have also been used to demonstrate sites of low current thresholds to electrical stimulation. However, AD thresholds have not held a consistent relationship with the ability to identify the site of spontaneous seizure onset (13). Also, they may demonstrate variability, with some thresholds being higher and some lower in the region of the epileptogenic zone. Because ADs probably require a certain neuronal density, in areas of significant neuronal loss (i.e., hippocampal sclerosis), thresholds may be higher in the region of epileptogenicity owing to the greater stimuli needed to recruit a smaller cellular population. Similarly, the presence of a foreign tissue lesion also makes the determination of the site of epileptogenicity using AD thresholds unreliable. Hence, AD thresholds have not been able to consistently identify epileptogenic and nonepiletpogenic tissue.

Technique Like ECOG, electrical stimulation for functional brain mapping can be performed in and out of the operating room. Electrical stimulation studies are usually performed in patients with indwelling electrodes following epilepsy monitoring. Sessions usually last anywhere from 30 minutes to several hours and may require more than one day to complete a functional map (Figure 17.7). Testing usually requires patients to be awake in order to obtain sensory, motor, visual, and language information. This may be tiring for the patient, although it is usually not painful. Stimulations are performed sequentially through pairs of adjacent subdural grid or strip contacts, with EEGs recorded simultane-

Husain 17

294

1/17/08

•

12:07 PM

Page 294

SE C T I ON I I : C l i n i c a l Me thod s

FIGURE 17.7 Working functional brain map of ECOG-guided electrical stimulation with a 4 by 5 contact subdural grid and a cavity wall 1 by 4 subdural strip placed over the central sulcus and inside an cystic cavity. Stimulation is repeated at each site and the duration and spread of the ADs as well as the clinical correlate is noted. Note the variability of stimulation, with the second inferior grid contact not reproducing sensory function and the third superior grid contact with sensory and motor function demonstrated during the second trial of stimulation.

ously from other contacts. Stimulation is with constant-current bipolar square-wave pulses delivered at 50 Hz for a duration of 0.5 ms. Initial stimulation intensity is 1 to 2 mA and is increased by 0.5 to 1 mA up to maximal settings allowable for the individual stimulator utilized (i.e., 10–15 mAs) and depending on the surrounding AD thresholds. Stimulation is continued for 4 to 5 seconds, with an AD defined as periodic epileptiform discharges or rhythmic epileptiform activity lasting at least 1 second in at least one contact. Two trials are performed to verify function, AD threshold without clinical symptoms, or no response at the maximum of 10 mA. When one trial results in an AD and another not, the tissue is regarded as noneloquent as long as no symp-

toms are registered during the trial with the AD. Repeat attempts to electrically stimulate sites of the cortex in gradually increasing intensities will often elicit an AD. During stimulation, stimulus artifact obscures the ECOG at the site of stimulation as well as adjacent electrodes (Figure 17.8). ADs are similar to restricted focal seizures and often consist of a sudden onset of recurrent rhythmic or epileptiform discharges that last for seconds to several minutes or more and can involve neighboring contacts. They may cease abruptly, have postictal suppression in the affected contacts, or evolve into a clinical seizure. Different locations within the brain have variable stimulus thresholds. Typically, mesial temporal stimulation has the

Husain 17

1/17/08

12:07 PM

Page 295

CH A P T ER 17: Epile ps y Surge r y

•

2 95

FIGURE 17.8 AD present at G5 electrode with some involvement of G8 and G9 on an implanted 20 contact subdural grid following 3.5 seconds of stimulation. Note the thick arrow at the onset of stimulus artifact and the thin arrow during the AD. LAT = left anterior temporal; LMT = left midtemporal; RT = right temporal.

lowest threshold for inducing ADs. However, there is a wide threshold range for eliciting ADs within the same cortical location of the same patient. In one study, up to a 9-mA difference was found within the temporal lobe and up to a 5.5-mA difference between adjacent electrodes on the temporal lobe (13). Furthermore, AD thresholds do not necessarily parallel a functional threshold, although this may vary from one site to another in an individual patient. Differences may even be seen from trial to trial or from day to day, making precise mapping even more complicated for precise demarcation of functions. A delay in responses to a single pulsed stimulation may reflect a higher risk of epileptogenicity than those appearing within 100 ms. The ECOG should always be monitored during electrical stimulation to assess the presence of ADs (Figure 17.8). If a seizure occurs with stimulation, the stimulus intensity is reduced and slowly retitrated up. In some cases this allows the current to be advanced to

higher or even maximal threshold intensity. Alternatively, the use of benzodiazepines may reduce the likelihood of an AD. When an AD occurs with postictal slowing, return of the EEG to baseline is required before continuing functional mapping. If a clinical seizure is precipitated by the stimulation, notation as to the habitual nature is documented and return of the EEG and the patient to baseline is required prior to continuing mapping. If lorazepam is required for a seizure lasting longer than 5 minutes, the session is usually aborted until a later time. Fortunately, seizures occur rarely during mapping sessions.

Other Forms of Functional Brain Mapping ECOG and electrical stimulation of the brain for functional mapping may be performed in conjunction with other techniques that help localize areas of eloquent cortex. Intraoperative median nerve somatosensory

Husain 17

296

1/17/08

•

12:07 PM

Page 296

SE C T I ON I I : C l i n i c a l Me thod s

evoked potentials (SEPs) are commonly obtained when identification of motor cortex is necessary. To record these potentials, a subdural grid or strip electrode is placed on the exposed cortex. The contralateral median nerve is stimulated. Cortical waveforms of the SEPs are obtained from the various electrode contacts referenced to a distal electrode, such as one on the contralateral mastoid. The N20 waveform is seen over the somatosensory cortex while a P22 (sometimes called the P20) waveform is seen over the motor cortex. The central sulcus lies between the electrodes generating the highest amplitude N20 to P22 complex and is easily delineated by the phase reversal that is generated (14) (Figure 17.9). This is not a true phase reversal, since the negative and positive waveforms do not represent two ends of a single dipole; rather, they represent two different dipoles. Hence this is often called a pseudo–phase reversal.

Several other nonneurophysiologic techniques are also used for brain mapping. The intracarotid amobarbital (Wada) test is used to lateralize language and memory; it is used most often in surgeries of the mesial temporal lobe. Functional magnetic resonance imaging (MRI) and positron emission tomography (PET) have shown utility in localizing motor, sensory, and language functions. Optical imaging is being used to map the spatiotemporal relationship between excitatory and inhibitory neuronal activity during epilepsy surgery. Neurophysiologic mapping with source and frequency analyses with magnetoencephalography (MEG) dipole analysis provides unique information not obtainable with anatomic or functional neuroimaging. Specialized surgical navigational systems are being developed that will coregister anatomic, functional, and neurophysiologic data to allow more precise delineation of the anatomic–functional boundaries of lesions.

FIGURE 17.9 Localization of the central sulcus using median nerve SEPs. Note the pseudophase reversal of the N20 (thick arrow) and P22 (thin arrow) waveforms at contacts 7 and 8 of the grid. The last channel demonstrates Erb’s point’s potential, verifying adequacy of stimulation. [Reproduced with permission from Husain (14).]

Husain 17

1/17/08

12:07 PM

Page 297

CH A P T ER 17: Epile ps y Surge r y

ANESTHESIA General anesthesia may produce significant alteration of the EEG background activity as well as the provoking epileptiform activity on the ECO General anesthesics are minimized or discontinued prior to intraoperative ECOG. However, they are a common choice during the induction phase of anesthesia for epilepsy surgery. Opioid analgesics such as fentanyl and remifentanil may activate IEDs and produce behavioral seizures. As noted previously, spike discharges may also be augmented or activated with methohexital. On the other hand, the use of benzodiazepines may accentuate both faster and slower frequencies on ECOG but act as suppressants for IEDs and seizures. Intravenous benzodiazepines remain the principal agents for aborting seizures both in and outside of the operating room. Propofol, etomidate, and barbiturates are general anesthetics that have dose-dependent effects upon ECOG, initially accentuating faster frequencies and then augmenting slower frequencies at higher doses. Inhalation halogenated anesthetics such as halothane, enflurane, and sevoflurane can influence or suppress background activity and alter epileptiform discharges. At higher concentrations, generalized delta frequencies become apparent and even a burst-suppression pattern may be noted. Enflurane may produce IEDs in nonepileptic patients or broaden the field of spread in those with seizures (2). Nitrous oxide can also suppress IEDs, and in addition, because of slowing introduced into ECOG, it is often discontinued prior to recording (6). Substituting nitrous oxide with intravenous anesthesia in concert with paralytic agents may minimize the effects on the recording. Local anesthesia is utilized when intact cognition is desirable for functional mapping. This is especially useful when regions close to eloquent cortex are to be excised. However, local anesthesia requires stable head position, cooperation from the patient, is more time consuming, and is difficult in the cognitively

•

2 97

challenged and younger age group. Even when local anesthesia is used for functional mapping, general anesthetics are used during the remainder of the surgery. Patients are awakened (transitioned to local anesthetic) when mapping is to be performed. Once mapping is complete, they are transitioned back to general anesthesia.

SAFETY Electrical safety of the brain is dependent on the charge density delivered to cortical tissue. Charge density [in microcoulombs (µC) of charge per square centimeter of tissue per phase of stimulation] is a measure of the amount of current applied during electrical stimulation. At the routine extraoperative settings utilized for functional mapping, no alteration in parenchymal architecture has been found with light microscopy at the sites of subdural electrode stimulation (15). Furthermore, no evidence of cumulative histologic injury at the electrode sites has been shown in any clinical trial using cortical or deep brain stimulation. Extraoperative stimulation using subdural electrodes may produce a charge density of 50 to 60 µC per square centimeter per phase (15). However, intraoperative stimulation can produce a much higher charge density (15,16). Repeated trials of stimulation have not demonstrated evidence of secondary epileptogenesis (i.e., kindling) in humans. Animal models, however, have demonstrated this phenomenon. It is thought that the human neocortex is much more difficult to kindle, and the irregular and limited electrical stimulation used for mapping is insufficient to produce this phenomenon. Morbidity may be associated with placement of depth and subdural electrodes. Cerebral edema with shifting of midline structures and infection are the primary concerns. Postoperative computed tomography (CT) of the brain and perioperative antibiotics are

Husain 17

298

1/17/08

•

12:07 PM

Page 298

SE C T I ON I I : C l i n i c a l Me thod s

routinely used to detect and prevent these complications. Presence of purulent wound drainage, unrelenting fever, disproportional change in mental status, or a notable increase in seizure frequency or status epilepticus should raise suspicion for complications.

TECHNICAL ISSUES Preparation Epilepsy surgery is different than other types of surgeries where NIOM is used. Rather than a standard procedure, in epilepsy surgery the procedure is individualized for the patient based on preoperative and intraoperative neurophysiologic findings. Prior to surgery, the NIOM team must be aware of the preoperative workup and the type of surgery planned. A discussion between the surgeon and neurophysiologist should establish whether ECOG, brain mapping, or both are to be performed. This must also be communicated to the anesthesiologist so that appropriate anesthetics are used. The NIOM team should discuss the monitoring that will be done during the procedure with the patient. The site of the incision and the amount of hair that will need to be shaved should be reviewed. During this time any potential issues that may hamper full patient cooperation are brought to the attention of the surgeon. An example of such an issue may be the need for the patient to have his or her glasses available during language mapping. Potential complications, especially in cases in which electrodes will be implanted for extraoperative recording, should be discussed with the patient. The NIOM technologist must make certain that appropriate equipment is available for the surgery. If ECOG is to be performed in the operating room, an EEG machine with the capacity to record 16 to 128 channels must be available. Newer portable digital EEG machines usually have this capability. If func-

tional mapping is to be performed, an electrical stimulator should also be available. A common type of stimulator is the Ojemann electrical stimulator which has a ball-tipped, handheld stimulating probe. A discussion between the surgeon and the NIOM team should establish the type of electrodes that will be used during the procedure and where they will be placed. This may include subdural grid or strip or other types of electrode. The size of these electrodes should also be determined. A few extra electrodes should be available in case the electrodes malfunction or are accidentally contaminated during surgery. All electrodes and stimulating probe should be sterilized prior to surgery.

Procedure Unlike other types of NIOM, a lot of patient preparation for monitoring is not needed after induction of anesthesia. If necessary, a scalp electrode contralateral to the side of surgery should be glued on with collodion for use a reference electrode. If median nerve SEPs are to be obtained, stimulating electrodes along the median nerve in the forearm and recording electrodes at Erb’s point should be placed. The rest of the electrodes will be placed by the surgeon after exposure. The electrodes to be used should be handed to the scrub nurse using sterile technique. Once surgical exposure is complete, the surgeon places the electrodes in predetermined locations. Once the electrodes are in position, the neurophysiologist with the help of the surgeon maps the location of the electrodes on a cartoon of the brain. This helps with identifying the location of abnormal discharges on the ECOG and mapping. Leads from the electrodes are passed to the NIOM technologist who inserts the pins into the jack box. Care must be taken to ensure that leads are inserted into the correct positions. A mistake in this can cause misinterpretation of the site of abnormal discharges. ECOG is recorded for several minutes and feedback regarding the presence of

Husain 17

1/17/08

12:07 PM

Page 299

CH A P T ER 17: Epile ps y Surge r y

spikes is given the by the neurophysiologist to the surgeon. The electrodes are then removed from the surgical field and resective surgery begins. In some situations the surgeon may request ECOG again after partial or complete resection. The virtues of postexcision ECOG have been discussed previously. After the initial ECOG, the surgeon may stimulate various cortical regions to determine AD thresholds. As noted previously, the surgeon stimulates initially with low intensity and then gradually increases the intensity. When an AD is noted, the NIOM team alerts the surgeon. When the ADs stop, the surgeon is again informed. If brain mapping to reproduce symptoms is to be performed, the anesthesiologist is informed and anesthesia is altered to awaken the patient. Motor and sensory mapping is performed by observing and talking with the patient during electrical stimulation of various parts of the cortex. Language mapping involves presenting pictures to the patient and asking him or her to name objects while different brain regions are stimulated by the surgeon. During this mapping, ECOG is monitored to note for ADs. If they occur, the surgeon is immediately notified. Median nerve SEPs are obtained if localization of the motor cortex is to be performed. Parameters for stimulation of the median nerve are the same as for other types of median nerve SEPs. The Erb’s point response is recorded as a measure of adequacy of stimulation. Cortical responses are recorded from the grid or strip electrodes placed on the surface of the brain. Usually a referential montage is used, with each subdural contact referenced to a distal electrode, usually on the contralateral mastoid. The grid or strip electrode is moved to different locations, and the site where the N20/P22 complex is of highest amplitude is noted. When subdural or depth electrodes are implanted for extraoperative recording, the NIOM team’s primary role in the operating room is to clearly identify which electrode is

•

2 99

in which location. This enables correct labeling of channels once the electrodes are connected to the EEG machine. After surgery, the patient undergoes x- ray andCT of the head to determine the sites of the electrodes. As soon as possible thereafter, the patient should be connected to the EEG machine and data acquisition started.

Postprocedure After surgery the electrodes are disposed in an appropriate container. If median nerve stimulation electrodes were used, they are removed and disposed of. A report documenting the findings is created as part of the patient’s permanent medical record. If extraoperative ECOG monitoring will be performed, the NIOM technologist places the head wrap as soon as the sterile dressings are placed. This is also done under the direct supervision of the circulating nurse or anesthesia team. Intracranial electrode leads are cleaned with saline or alcohol, dried, and attached to the recording cables, which will be maintained outside the sterile field. Surgical drains that have been placed may be moved to one side so that they can be removed safely at the bedside by the neurosurgeon without strain on the electrode leads. Once the patient arrives in the epilepsy monitoring or intensive care unit, the leads are connected to the jack box of the EEG machine. Monitoring is continued until an adequate number of seizures are recorded and all extraoperative functional mapping is completed. Thereafter the patient returns to the operating room for removal of the electrodes. The NIOM team is usually not needed during this procedure.

CONCLUSIONS Epilepsy remains the prime target for using ECOG. The application of ECOG for patients with known structural lesions noted on brain MRI has been controversial, and

Husain 17

300

1/17/08

•

12:07 PM

Page 300

SE C T I ON I I : C l i n i c a l Me thod s

resection of the lesion is usually performed with excellent results. Most of the ECOGguided reports have been associated with temporal lobe resections in adults with epilepsy and have been retrospective in nature. The utility of preexcision ECOG to tailor resections has been inconsistent. Similarly, the prognostic value of postexcision ECOG has also been mixed, with outcomes contingent on the margins of the resection. Unfortunately, IEDs found on ECOG have not demonstrated consistent results, and “spike chasing” during intraoperative ECOG has not yet proven to be of value. Therefore the use of ECOG in isolation has not received validation to recommend its routine use to extend cortical resections, although questions still remain regarding its utility in an individual patient. Additionally, the potential benefits of using ECOG in patients without seizures, newly diagnosed seizures, or controlled epilepsy are still to be defined. Electrical brain stimulation for functional brain mapping is routinely used in conjunction with ECOG. It provides a map of eloquent cortex that can guide surgical resection. ECOG and electrical stimulation of the brain are potentially beneficial and have been applied toward helping patients with medically intractable seizures undergoing epilepsy surgery.

REFERENCES 1. Chatrian G-E. Intraoperative electrocorticography. In: Ebersole JS, Pedley TA, eds. Current Practice of Clinical Electroencephalography, 3rd ed. Philadelphia: Lippincott Williams & Wilkins, 2003:681–712. 2. Luciano D, Devinsky O, Pannizzo F. Electrocorticography during cortical stimulation. In: Devinski O, Beric A, Dogali M, eds. Electrical and Magnetic Stimulation of the Brain and Spinal Cord. New York: Raven Press, 1993:87–102. 3. Engel J Jr, Driver MV, Falconer MA. Electrophysiological correlates of pathology

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

and surgical results in temporal lobe epilepsy. Brain 1975;98:129–156. Tran TA, Spencer SS, Marks D, et al. Significance of spikes recorded on electrocorticography in nonlesional medial temporal lobe epilepsy. Ann Neurol 1995;38:763–770. Keene DL, Whiting S, Ventureyra EC. Electrocorticography. Epileptic Disord 2000; 2:57–63. Binnie CD, Polkey CE, Alarcon G. Electrocorticography. In: Luders HO, Comair YG, eds. Epilepsy Surgery. Philadelphia: Lippincott Williams & Wilkins, 2001:637–641. Berger MS, Ghatan S, Haglund MM, et al. Low-grade gliomas associated with intractable epilepsy: seizure outcome utilizing electrocorticography during tumor resection. J Neurosurg 1993;79:62–69. Ferrier CH, Aronica E, Leijten FS, et al. Electrocorticographic discharge patterns in glioneuronal tumors and focal cortical dysplasia. Epilepsia 2006;47:1477–1486. Tran TA, Spencer SS, Javidan M, et al. Significance of spikes recorded on intraoperative electrocorticography in patients with brain tumor and epilepsy. Epilepsia 1997;38: 1132–1139. Leijten FS, Alpherts WC, Van Huffelen AC, et al. The effects on cognitive performance of tailored resection in surgery for nonlesional mesiotemporal lobe epilepsy. Epilepsia 2005; 46:431–439. Fish DR, Gloor P, Quesney FL, Olivier A. Clinical responses to electrical brain stimulation of the temporal and frontal lobes in patients with epilepsy. Pathophysiological implications. Brain 1993;116(Pt 2):397–414. Branco DM, Coelho TM, Branco BM, , et al. Functional variability of the human cortical motor map: electrical stimulation findings in perirolandic epilepsy surgery. J Clin Neurophysiol 2003;20:17–25. Lesser RP, Luders H, Klem G, et al. Cortical afterdischarge and functional response thresholds: results of extraoperative testing. Epilepsia 1984;25:615–621. Husain AM. Neurophysiologic intraoperative monitoring. In: Tatum WO, Husain AM, Benbadis SR, Kaplan PW, eds. Handbook of EEG Interpretation. New York: Demos, 2007:223–260.

Husain 17

1/17/08

12:07 PM

Page 301

CH A P T ER 17: Epile ps y Surge r y 15. Gordon B, Lesser RP, Rance NE, et al. Parameters for direct cortical electrical stimulation in the human: histopathologic confirmation. Electroencephalogr Clin Neurophysiol 1990;75:371–377.

•

3 01

16. Pouratian N, Cannestra AF, Bookheimer SY, et al. Variability of intraoperative electrocortical stimulation mapping parameters across and within individuals. J Neurosurg 2004;101(3): 458–466.

Husain 17

1/17/08

12:07 PM

Page 302

Husain 18 Index

1/17/08

12:09 PM

Page 303

Index

Note: Bold numbers indicate illustrations, italic t indicates a table.

desflurane in, 58, 58t dexmedetomidine in, 63 electrocorticography (ECOG) and, 297 electroencephalography (EEG) and, 56–60, 59t, 239–240, 272–273, 273, 278–279, 297 electromyography (EMG) and, 28–29, 172, 173, 205 enflurane in, 58, 58t epilepsy/epilepsy surgery and, 297 etomidate in, 60, 60t, 62–63, 130 equipment used for, 11, 11 evoked potentials (EPs) and, 58, 58t, 60, 60t extradural, 119–121, 119t extramedullary, 119–121, 119t fast anterior dominant rhythmic activity (FAR) in, 272, 273 frontal intermittent rhythmic delta activity (FIRDA) in, 272 gamma aminobutyric acid (GABA) and, 61 general, 55 halogenated agents in, 59 halothane in, 58, 58t hematology and, 57 hyper- and hypotension and, 57 induction of, 55 inhalation agents used for, 58, 58t, 59t, 130 intracranial pressure and, 57–58 intramedullary, 119–121, 119t intravenous agents in, 60–63, 60t isoflurane in, 58, 58t ketamine in, 60, 60t, 62, 130 local, 55 mean/minimal alveolar concentration (MAC) and, 55–56, 58, 59, 62, 63, 64, 130, 239, 272 motor evoked potentials (MEPs) and, 37–38, 38t, 59, 60, 129–130, 239 nerve action potentials (NAPs) and, 204 neuromuscular blockade type, 130, 147–148, 148, 149

access to patient, 15, 16–17 achondroplasia, 97 ACNS. See American College of Neurosurgeons afferent pathway of spinal cord, 118–119 afterdischarge (AD) localization, 293 alfentanil, 130 American Board of Electroencephalographic and Evoked Potential Technologists (ABRET), 70–71 American College of Neurosurgeons (ACNS), 73, 78, 81, 100 American Medical Association (AMA), 67 amplifiers and operational amplifiers, 73, 74, 79–80, 79, 80, 81t common-mode rejection ratio in, 80–81, 80 electroencephalography (EEG), 81 electromyography (EMG), 81 evoked potentials (EPs), 81 input impedance in, 81, 81 nerve conduction studies (NCS), 81 operational, 79–80, 79, 80, 81t process of, 81–82 voltage-divider rule in, 81, 81 amplitude of MUPs, 26 anal sphincter function monitoring, 161–162, 161 analog filters, 82–83 analog-to-digital conversion (ADC), 83–86, 85 anesthesia, 15, 55–66 agents used in, 55, 58–63, 58t, 59t anterior intermittent slow waves (AIS) in, 272 barbiturates in, 60, 60t, 61, 130 benzodiazepine in, 60, 60t, 61 blood flow and, 57 brainstem auditory evoked potentials (BAEPs) and, 43, 59, 64, 201 carotid endarterectomy (CEA) and, 272–273, 278 classification of, 119, 119t compound muscle action potentials (CMAPs) and, 205 303

Husain 18 Index

304

•

1/17/08

12:09 PM

Page 304

In d e x

anesthesia (continued) nitrous oxide in, 58, 58t, 60, 130 nonpharmacologic factors and, NIOM and, 56–58 opioids in, 60, 60t, 61–62, 130 paralytics in, 63, 64 potassium/sodium/calcium balance in, 58 principles of, 55–56 propofol in, 60, 60t, 61, 130 regional, 55 relative insensitivity of NIOM to, 64 sedation as, 55 selective dorsal rhizotomy (SDR) and, 172, 173 sensitivity to, 64–65 sevoflurane in, 58, 58t somatosensory evoked potentials (SEPs) and, 33, 57, 59, 60, 64, 128–130, 239–240, 244, 272–273, 278–279 spinal cord surgery and, 128–130 techniques for, 63–65 temperature and, 56–57 thoracic aortic surgery and, 239–240, 243–244, 246 total intravenous (TIVA), 60, 61, 62, 64, 128, 130, 243–244, 246 train-of-four (TOF) blockade and, 63, 147–148, 148, 149 ventilation and, 57 vertebral column surgery and, 112–113 widespread anteriorly maximum rhythmic activity (WAR) in, 272 widespread persistent slow waves (WPS) in, 272, 273 anesthesia team/technician/resident, 7 anesthesiologist, 7 aneurysmal bone cyst, 119–121, 119t aneurysms, 41, 229–231, 229, 230 anterior cervical discotomy and fusion, 98 anterior intermittent slow waves (AIS), 272 anterior spinal release surgery, 98 antiepileptic drugs (AEDs), 283 aorta/aortic diseases. See thoracic aortic surgery aortic dissection, 231–232, 231 arachnoid cysts, 215t. See also tumors of the cerebellopontine angle Arnold-Chiari malformation, 97 arteries of spinal cord, 119 arteriovenous malformations (AVM), 41, 117 artifacts, 74 automated rejection of, 83, 84 electroencephalography (EEG) and, 40 free-running EMG and, 142–143, 143 operating table, 13 ascending aorta and aortic arch, 232–233, 232, 233, 238–241, 249–254, 256. See also thoracic aortic surgery aseptic technique, 3–6 aspirator, Cavitron ultrasonic surgical (CUSA), 11 astrocytoma, 119–121, 119t asynchronous EMG pattern, 25–26 atherosclerosis formations, 263–264, 265

atracurium, 130 attendants, OR, 8 attending anesthesiologist, 7 attending surgeon, 6 attire for the operating room, 5, 17 auditory pathways, for brainstem auditory evoked potentials (BAEPs) and, 41–42 auditory stimulation, 77–79 piezoelectric transducers for, 78–79 sound and decibel scale in, 77–79 sound pressure level (SPL) and, 78 sound reference and decibel scale in, 78 types of, 78–79 autoclaves, 4 automatic artifact rejection, 83, 84 Axon Systems, 106 axonal injury and EMG activity, 26–27 bandpass filter, 82, 82 barbiturates, 60, 60t, 61, 130, 239, 297 Bell, Alexander Graham, 77 benzodiazepine in, 60, 60t, 61, 239, 297 Berger, Hans, 286 bilateral independent lateralized epileptiform discharges (BiPLEDs), 240 billing. See coding and billing biopsy, fascicle selection for, 190 bits and vertical resolution, 85–86 blood flow and NIOM, 57, 96 BrainLab, 12, 13 brainstem auditory evoked potentials (BAEP), 15, 21, 41–43 anesthesia and, 43, 59, 64, 201 auditory stimulation and, 41–42, 78, 79 electromyography (EMG) and, 43 factors affecting, 43 filtration in, 82 indications for, 43 interpretation of, 42–43, 201–203, 201, 202, 203 intracranial pressure and, 58 ketamine and, 62 microvascular surgery (MVD) and, 199–203, 200, 201, 202, 203, 206, 207, 209 operational amplifiers in, 80 paralytics in, 63, 64 preparation, procedure, postprocedure, 207–210, 221–226 recording techniques for, 42 stimulation techniques for, 42, 78, 79 tumors of the CPA and, 214, 216–220, 216, 221, 222 vascular surgeries and, 276–277 burns, 16 bursts, EMG, 23, 24 C-arm, 11, 12 Cadwell, 106 carbon dioxide (CO2) lasers, 11, 122 cardiopulmonary bypass (CPB), 8, 9, 9, 232, 233–234, 234, 237–238

Husain 18 Index

1/17/08

12:09 PM

Page 305

Ind e x carotid cross clamping (CCC) procedure, 270, 273–275 carotid endarterectomy (CEA), 261–281 alarm criteria during, 273–275, 275 anatomy and pathology in, 261–265, 262, 263, 264, 265 anesthesia and, 272–273, 278–279 atherosclerosis formations in, 263–264, 265 brainstem auditory evoked potentials (BAEPs) and, 276–277 carotid cross clamping (CCC) procedure in, 270, 273–275 cerebral blood flow (CBF) assessment during, 2, 269–270, 270, 273–275 clinical symptoms/presentation of atherosclerotic disease and, 265–267 collateral circulation in, 261–262, 263 electroencephalography (EEG) and, 39, 40, 41, 270–278, 276 electromyography (EMG) in, 279–280 evidence supporting NIOM use during, 277–278 hypoglossal nerve and, 262, 264 mean arterial pressure (MAP) and, 272 other vascular surgeries and, 276–277 quantitative EEG (QEEG) in, 271–276, 276, 278–280 raw EEG (REEG) in, 271–276, 274, 277–280 reversible ischemic neurologic deficit (RIND) in, 266 sample NIOM case in, 275–276 somatosensory evoked potentials (SEPs) in, 270–276, 277, 277–278 surgical procedure for, 267–269, 268 technical considerations in, preparation, procedure, postprocedure, 278–280 transcranial Doppler (TCD) in, 270 transient ischemic attack (TIA)/stroke and, 263, 266–267, 266 utility of NIOM during, 269–270 vagus nerve and, 262, 264 carotid surgery. See carotid endarterectomy carpal tunnel syndrome, 188, 189 Cavitron ultrasonic surgical aspirator (CUSA), 11, 122, 135 CD storage media, 88 cell savers, 12 cerebellopontine angle surgery, tumor. See tumors of the cerebellopontine angle cerebral blood flow (CBF) assessment, 269–270, 270, 273–275 cerebral palsy, 97, 169–170. See also selective dorsal rhizotomy cerebrospinal fluid (CSF) tests, 121, 242 cervical spondylosis, 97, 98 chaining of stimulators, 74. See also interleaving channels for EP machines, 73–74 chemical sterilization, 4 chondroblastoma, 119–121, 119t chondrosarcoma, 119–121, 119t chordoma, 119–121, 119t

•

3 05

circulating nurse, 8 clean-up of OR, 17 client-server systems, in remote monitoring, 51–52 clinical registered nurse anesthetist (CRNA), 7 CMAPs. See compound muscle action potentials Cobb angle, 96 Code 95829, 69 Code 95920, 68–69, 68t, 68 Code 95955, 69 Codes 95961/95962, 69 Codes 95970–95979, 69, 70t, 69 coding and billing, 67–69, 70t cortical and subcortical localization (Codes 95961/95962) in, 69 Current Procedural Coding (CPT) for billing, 67 electrocorticography (ECoG; Code 95829) in, 69 intraoperative neurophysiology CPT, 67, 68t intraoperative neurophysiology testing (Code 95920) in, 68–69, 68t neurostimulator programming and analysis (Codes 95970–95979) in, 69, 70t nonintracranial EEG (Code 95955 in, 69 Relative Value Units (RVU) in, 67 Resource Based Relative Value System (RBRVS) in, 67 collateral circulation in carotid artery, 261–262, 263 common peroneal neuropathy at the knee, 189 common-mode rejection ratio, op amps, 80–81 compound muscle action potential (CMAPs), 29, 125, 129–131, 133–135 anesthesia and, 205 interpretation of, 205–206 lumbosacral surgery and, 148, 149, 150, 152 microvascular surgery (MVD) and, 205–206 neuromuscular blockade and, 148, 149 peripheral nerve surgery and, 182, 183–188, 190–191 tethered cord syndrome (TCS) and, 158–159, 159 trigeminal nerve monitoring using, 220 tumors of the CPA and, 219, 224 compressed spectral array (CSA) EEG, 89 compression, spinal, 97 computed tomography (CT), 121, 158 computer system for NIOM machines, 83–89 in remote monitoring, 45–48 connecting to a network, for remote monitoring, 47–50 constant voltage vs. constant current, in stimulators, 76–77, 76 contradictions to NIOM, 14 conversion, analog-to-digital (ADC), 83–86, 85 cords and cables, safety, 15–16 cortical and subcortical localization (Codes 95961/95962), 69 cortical potentials, SEP, 100–104 corticospinal tracts, 96 Crawford classification of aortic aneurysms, 230, 230

Husain 18 Index

306

•

1/17/08

12:09 PM

Page 306

In d e x

credentialing of staff, 70–71 Current Procedural Coding (CPT) for billing, 67 cutoff frequency, filter, 83, 83 data displays, 88–89 data packet, in remote monitoring, 47–48, 48 data transmission, for remote monitoring, 50–54 DeBakey and Stanford classification of aortic dissection, 231, 231 decibel scale, 77–79 decompression, microvascular. See microvascular decompression decompression, posterior, 99 decussation (SEPs), 100–104, 101 deep hypothermic circulatory arrest (DHCA), 232–241, 234, 240, 241, 241t, 249–250 electroencephalography (EEG) and, 240–241, 240, 241, 241t somatosensory evoked potentials (SEPs) and, 241 default gateways for remote monitoring computers, 47 density spectral array (DSA) EEG, 89 dermatomal somatosensory evoked potentials (DSEP), 139, 140 descending thoracic and thoracoabdominal aorta, 233–235, 235, 241–245, 250–252, 254–257. See also thoracic aortic surgery desflurane, 58, 58t, 59 dexmedetomidine, 63 diastematomyelia (split-cord malformation), 97, 157. See also tethered cord syndrome diazepam, 170 digital filters, 87–88 Digitimer, 106 disinfection, 17 display software, in remote monitoring, 50–51 displays, 88 distal nerve disturbances, 110 documentation, 17–18, 71 dorsal column, 118 dorsal nerve rootlet, selective dorsal rhizotomy (SDR) and, 174–177, 175, 176 dorsal rhizotomy. See selective dorsal rhizotomy, 169 drapes for sterile field, 5–6 drill, surgical, 13 dural sinus thromboses, 215t. See also tumors of the cerebellopontine angle duration of stimulus, 75–76, 75 DVD storage media, 88 dwell time, 84–85 dysplasia, vertebral, 96 ECG. See electrocardiograph ECOG. See electrocorticography EEG. See electroencephalography efferent pathway of spinal cord, 118–119, 118 EKG. See electrocardiograph elbow, ulnar neuropathy at (UNE), 188 electrical safety, 15–16

electrocardiography (ECG/EKG) peripheral potentials in, 105 signal-to-noise ratio (SNR) and, 86 thoracic aortic surgery and, 251 electrocautery unit, 9, 9 electroconvulsive therapy (ECT), 61, 62 electrocorticography (ECOG), 69, 283, 286–291 anesthesia and, 297 background of, 286 electroencephalography (EEG) used with, 286, 287, 289 electrodes used in, 287–288, 287, 288 epilepsy/epilepsy surgery and, 283, 286–291 epileptiform abnormalities in, 289–290 evidence of usefulness of, 290–291 extraoperative recording using, 286–287 functional brain mapping and, 291–296 interictal epileptiform discharges (IEDs) and, 286, 289 interpretation of, 289–290 nonepileptiform abnormalities in, 290 postexcision, 291 preexcision, 290–291 preparation, procedure, postprocedure, 298–299 technique for, 288–289 temporal lobe epilepsy (TLE) and, 291 video-EEG and, 287, 289 electroencephalography (EEG), 9, 15, 21, 38–41, 39, 74, 249–250 amplification in, 81 anesthesia and, 56, 58, 59, 59t, 60, 239–240, 272–273, 273, 278–279, 297 aneurysms and, 41 anterior intermittent slow waves (AIS) in, 272 applications for, 41 arteriovenous malformations (AVMs) and, 41 artifacts in, 40, 83 benzodiazepines in, 61 bilateral independent lateralized epileptiform discharges (BiPLEDs) in, 240 carotid endarterectomies (CEAs) and, 39, 40, 41, 270–278, 276 changes of significance in, 41 coding and billing for, 67, 69 compressed spectral array (CSA), 89 data analysis of, 88 data displays for, 89 deep hypothermic circulatory arrest (DHCA) in, 240–241, 240, 241, 241t density spectral array (DSA), 89 digital signals in, 84 electrocorticography (ECOG) and, 286, 287, 289 epilepsy/epilepsy surgery and, 283–300, 285 etomidate in, 62–63 fast anterior dominant rhythmic activity (FAR) in, 272, 273 filtration in, 82 frequency classification of, 39 frontal intermittent rhythmic delta activity (FIRDA) in, 272

Husain 18 Index

1/17/08

12:09 PM

Page 307

Ind e x functional brain mapping and, 295 generalized periodic epileptiform discharges (GPEDs) and, 240, 240 generators for, 38–39 montages of recording in, 39, 39, 40 motor evoked potentials (MEPs) and, 40, 40 opioids in, 61–62 paralytics in, 63, 64 patterns seen in, 39–40 periodic lateralized epileptiform discharges (PLEDs), 240 preparation, procedure, postprocedure for, 251–257, 278–280, 298–299 processing methods for, 40–41 propofol in, 61 quantitative (QEEG), 271–276, 276, 278–280 raw (REEG), 271–276, 274, 277–280 signal-to-noise ratio (SNR) and, 86 somatosensory evoked potentials (SEPs) and, 41 temperature effects on, 56 tethered cord syndrome (TCS) and, 158 thoracic aortic surgery and, 239–240 tumors of the CPA and, 222 vertebral column surgery and, 101 video-, 287, 289 widespread anteriorly maximum rhythmic activity (WAR) in, 272 widespread persistent slow waves (WPS) in, 272, 273 electromagnetic interference (EMI), 177 electromyography (EMG), 15, 21, 74 amplification in, 81 amplitude of MUPs in, 26 anatomy, physiology and, 21, 23 anesthesia and, 28–29, 172, 173, 205 asynchronous pattern in, 25–26 brainstem auditory evoked potentials (BAEPs) and, 43 bursts and trains in, 23, 24 carotid endarterectomy (CEA) and, 279–280 characterization of, 23 compound muscle action potentials (CMAPs) and, 150, 152 data analysis of, 88 data displays for, 88–89 digital signals in, 84 facial nerve monitoring using, 219 filtration in, 82 firing rate of MUPs in, 25–26 free-running, 21–29, 139–143, 140 hemifacial spasm (HFS) and, 196 interpretation of, 23, 205–206 ischemia and, 27 lumbosacral surgery and, 139–154, 140 microvascular surgery (MVD) and, 199, 205–207 motor nerve conduction studies (triggered) using, 27–29 motor unit potential (MUP) in, 21, 23, 25–26 muscle selection for, 21, 23t, 152–153

•

3 07

myotomes and, 21 needle electrodes for, 186, 186 nerve dysfunction vs. activity on, relation of, 26–27 neuromuscular blockade monitoring using, 147–148, 148, 149, 153 parameters for, 150t patient safety and, 153 pedicle holes and screws evaluation by, 27–28, 28t, 29, 144–147, 144, 145, 146t peripheral nerve surgery and, 186–187 preparation, procedure, postprocedure, 177–180, 190–191, 207–210, 221–226 radiculopathy in, 110 recording and stimulating parameters used in, 21, 22t recording electrodes for, 150, 150t, 151 reflexive activity and, 25 sampling frequency for, 85 selective dorsal rhizotomy (SDR) and, 171–172, 173 shape of MUPs in, 26 signal-to-noise ratio (SNR) and, 86 “significant” activity on, communicating with the surgeon, 27 sphincter function monitoring using, 161–162, 161 spinal cord surgery and, 136 stimulating electrodes and parameters for, 75, 151, 152, 152 subdermal needle electrodes used in, 173 surgical events and, relationship between, 24–25, 26 tethered cord syndrome (TCS) and, 158–163, 160 threshold ratios (TR) and, 173–174, 175t trigeminal nerve monitoring using, 219–220 tumors of the CPA and, 217–220, 221, 225 EMG. See electromyography endovascular stent grafts (EVSG), 235–237, 236, 237, 246–251 endovascular thoracic aortic repair, 235–237, 236, 237, 246–251. See also thoracic aortic surgery enflurane, 58, 58t, 59, 297 EPs. See evoked potentials ependymoma, 119–121, 119t epidermoids, 215t. See also tumors of the cerebellopontine angle epilepsy/epilepsy surgery, 283–301 afterdischarge (AD) localization in, 293–295, 295 anatomy and physiology in, 283–284 anesthesia and, 297 antiepileptic drugs (AEDs) and, 283 bilateral independent lateralized epileptiform discharges (BiPLEDs) in, 240 electroencephalography (EEG) and, 283–300, 285, 295 electrocorticography (ECOG) and, 283, 286–291 functional brain mapping and, 283, 291–296

Husain 18 Index

308

•

1/17/08

12:09 PM

Page 308

In d e x

epilepsy/epilepsy surgery (continued) generalized periodic epileptiform discharges (GPEDs), 240, 240 interictal epileptiform discharge (IED) in, 284, 285 intracarotid amobarbital (Wada) test and, 296 localization-related (LRE), 283 magnetic resonance imaging (MRI) and, 296 magnetoencephalography (MEG) and, 296 pathology in, 284–286 periodic lateralized epileptiform discharges (PLEDs), 240 positron emission tomography (PET) and, 296 safety precautions for, 297–298 somatosensory evoked potentials (SEPs) in, 299 technical considerations in, preparation, procedure, postprocedure, 298–299 temporal lobe (TLE), 284, 285, 291 video-EEG and, 287, 289 equipment for NIOM, 73–92 equipment for the OR, 8–13 Erb’s point, 104 errors in medical care, 71 Ethernet connector, 47, 47 ethical responsibility of staff, 70–71 etiquette of the OR, 16–17 etomidate in, 60, 60t, 62–63, 130, 297 evoked potentials (EPs), 73 amplification in, 81 anesthesia and, 58, 58t, 60, 60t artifact rejection in, 83 auditory stimulation for, 78 data displays for, 88–89 digital signals in, 84 distal nerve disturbances in, 110 interpretation of, 108–111 ketamine and, 62 opioids in, 61–62 pathologic decrements in, 110, 113, 114 peripheral potentials in, 105 preparation, procedure, postprocedure for, 111–115 proximal nerve or plexus disturbances in, 110 radiculopathy in, 110 signal averaging and, 86–87 signal-to-noise ratio (SNR) and, 86, 87 spinal cord compromise in, 110–111, 111 stimulators and, 75, 78 systemic factors affecting, 109 technical factors affecting, 109–110 temperature effects on, 56 tumors of the CPA and, 219 vertebral column surgery and, 95 warning criteria for, 108–111, 109 waveform of, 86–87, 87 Ewing’s sarcoma, 119–121, 119t extradural spinal tumors, 119–121, 119t extramedullary spinal tumors, 119–121, 119t facial nerve monitoring, 219 Faraday cage, 73

fascicle selection for nerve biopsy, 190 fasciculus gracilis, 118 fasciculus cuneatus, 118 fast anterior dominant rhythmic activity (FAR), 272, 273 feedback rapidity, in SEPs, 100 fentanyl, 297 fiber tracts of spinal cord in, 118–119 filtering data, in remote monitoring computers, 47–48 filters, 74, 82–83, 82, 83 amplitude vs. frequency in, 82–83, 82 analog, 82–83 bandpass, 82, 82 cutoff frequency of, 83, 83 digital, 87–88 high-cut, 82–83 low-cut, 82–83 order of, 82–83 roll-off of, 82–83 firewalls, in remote monitoring computers, 47, 48, 50, 51 53, 53 firing rate of MUPs, 25–26 fluoroscopy, 11, 16 follow-up, 17 fractures, spinal, 97, 98 Free Software Foundation, 52 free-running EMG, 21, 139. See also electromyography artifacts in, 142–143, 143 muscles suitable for, 141, 141t outcome data from, postop, 141, 142t output from, 140–141, 141 precautions for, 142–143 radiculopathy and, 142–143, 143t, 143 sensitivity of, 141–142, 142t technique for, 140–141 tethered cord syndrome (TCS) and, 161, 162 Friedreich’s ataxia, 97 frequency, electroencephalography (EEG) and, 39 frontal intermittent rhythmic delta activity (FIRDA), 272 fully qualified names, for remote monitoring and computers, 51 functional brain mapping, 283, 291–296 afterdischarge (AD) localization using, 293–295, 295 electroencephalography (EEG) and, 295 epilepsy/epilepsy surgery and, 283, 291–296 intracarotid amobarbital (Wada) test and, 296 localization using, 292–293, 296 magnetic resonance imaging (MRI) and, 296 magnetoencephalography (MEG) and, 296 positron emission tomography (PET) and, 296 preparation, procedure, postprocedure, 298–299 somatosensory evoked potentials (SEPs) and, 296–297 technique for, 293–295, 294, 295 utility of, 291–292

Husain 18 Index

1/17/08

12:09 PM

Page 309

Ind e x future directions tethered cord syndrome (TCS) and, 166–167 thoracic aortic surgery and, 257 gamma aminobutyric acid (GABA), 61 ganglioglioma, 119–121, 119t general anesthesia. See anesthesia General Public License, 52 generalized periodic epileptiform discharges (GPEDs), 240, 240 generators, for somatosensory evoked potentials (SEPs), 30–31, 30t giant–cell tumor, 119–121, 119t globus tumors, 215t. See also tumors of the cerebellopontine angle glossopharyngeal neuralgia (GPN), 195, 197, 199. See also microvascular decompression GoToMyPC, remote monitoring and, 52 gowns, masks, drapes, 5–6 grounding, 16 halogenated agents in anesthesia, 59. See also anesthesia halothane in, 58, 58t, 59, 297 Health Insurance Portability and Accountability Act (HIPAA), 53, 88 heat sterilization, 3–4 hemangioblastoma, 119–121, 119t hemangioma, 119–121, 119t hematology, anesthesia and, 57 hemifacial spasm (HFS), 195, 196–199. See also microvascular decompression hemivertebrae, 96 high-cut filters, 82–83 history of NIOM technology, 73–74 horizontal gaze palsy and progressive scoliosis (HGPPS), 97 horizontal resolution, 84–85 hypertension, 57 hypocarbia, 57 hypoglossal nerve, 262, 264 hypotension, 57, 110 hypoxia, 57 image capture, for remote monitoring, 53–54 image-guided surgery systems, 12, 13 infection control, 16 infratentorial lateral supracerebellar approach (ILSA), 197, 198 inhalation agents used for anesthesia, 58, 58t, 59t injectors, fluoroscopy, 12 Inomed, 106 input impedance, amplifiers, 81, 81 Institute of Medicine, 71 Institutional Review Board (IRB), 77 intensity of stimulus, 75, 75 interictal epileptiform discharge (IED), 284, 285, 286, 289, 291 interleaving, 74 internal carotid artery (ICA). See carotid endarterectomy (CEA)

•

3 09

Internet and remote monitoring, 45–48 interpeak latency (IPL), 217 interpreting BAEP, 42–43, 201–203, 201, 202, 203 compound muscle action potentials (CMAPs), 205–206 electrocorticograph (ECOG), 289–290 electromyography (EMG), 23, 205–206 muscle evoked potential (MEP), transcranial, 35–36 nerve action potentials (NAPs), 204–205 somatosensory evoked potentials (SEPs), 31–33 intracarotid amobarbital (Wada) test, 296 intracranial pressure, anesthesia and, 57–58 intramedullary spinal tumors, 119–121, 119t intraoperative neurophysiology CPT codes, 67, 68t intraoperative neurophysiology testing (Code 95920) in, 68–69, 68t intraoperative ultrasound, 10 intrathecal pumps, spasticity and, 170 intravenous agents for anesthesia, 60–63, 60t inverting op amp, 79–80, 79, 80, 81t IP (Internet Protocol) addresses, in remote monitoring computers, 45, 48, 51 ipconfig command, in remote monitoring computer systems, 45, 47 ischemia, electromyography (EMG) and, 27 isoflurane in, 58, 58t, 59 keep-alive signals, in remote monitoring, 52–53, 53 ketamine, 60, 60t, 62, 130 knee, common peroneal neuropathy at, 189 laptop NIOM machines, 74 lasers, 11, 122 lateral suboccipital infrafloccular approach (LSIA), 197 legal issues, 71 leiomyosarcoma, 119–121, 119t liability for errors, 71 lipoma, 119–121, 119t, 215t. See also tumors of the cerebellopontine angle local anesthesia. See anesthesia localization, somatosensory evoked potentials (SEPs) and, 30–31, 31t localization-related epilepsy (LRE), 283 lorazepam, 170 Louisiana State University, 188 low-cut filters, 82–83 lower limb derivations (SEP), 101–104, 102, 103, 104t lumbosacral surgery, 139–154 compound muscle action potentials (CMAPs) in, 148, 149, 150, 152 dermatomal somatosensory evoked potentials (DSEP) in, 139, 140 free-running EMG in, 139–143, 140 artifacts in, 142–143, 143 muscles suitable for, 141, 141t outcome data from, postop, 141, 142t output from, 140–141, 141

Husain 18 Index

310

•

1/17/08

12:09 PM

Page 310

In d e x

lumbosacral surgery (continued) precautions for, 142–143 radiculopathy and, 142–143, 143t, 143 sensitivity of, 141–142, 142t technique for, 140–141 MEPs in, 139 muscle selection for EMG in, 152–153 neuromuscular blockade monitoring in, using EMG, 147–148, 148, 149, 153 patient safety and, 153 pedicle hole/screw evaluation in, using EMG, 144–147, 144, 145, 146t preoperative studies for, 150 recording electrodes and parameters for, 150, 150t, 151 somatosensory evoked potentials (SEPs) in, 139, 140 stimulating electrodes and parameters for, 151, 152, 152 magnetic resonance imaging (MRI), 12, 121 epilepsy/epilepsy surgery and, 296 hemifacial spasm (HFS) and, 197 peripheral nerve surgery and, 190 tethered cord syndrome (TCS) and, 157 trigeminal neuralgia (TN) and, 196 tumors of the CPA and, 215 magnetoencephalography (MEG), 296 masks, 5–6 mean/minimum alveolar concentration (MAC), 55–56, 58, 59, 62–64, 130, 239 mean arterial pressure (MAP), 243, 272 median neuropathy at the wrist (carpal tunnel syndrome) and, 188, 189 medical students, 6–7 medical-legal issues, 71 meningioma, 119–121, 119t. See also tumors of the cerebellopontine angle MEPs. See motor evoked potentials methohexital, 297 microscope, 9–10, 10 microvascular decompression (MVD), 195–210 brainstem auditory evoked potentials (BAEPs) in, 199–203, 200, 201, 202, 203, 206, 207, 209 compound muscle action potentials (CMAPs) in, 205–206 electromyography (EMG) in, 199, 205–207 in glossopharyngeal neuralgia (GPN), 195, 197, 199 in hemifacial spasm (HFS), 195, 196–199 infratentorial lateral supracerebellar approach (ILSA) in, 197, 198 lateral suboccipital infrafloccular approach (LSIA) in, 197 microvascular decompression; MVD. See microvascular decompression (MVD), 195 monitoring techniques used in, 199–206 nerve action potential (NAPs), vestibulocochlear nerve in, 204–205, 204, 206 nerve conduction studies (NCS) in, 207

somatosensory evoked potentials (SEPs) in, 199, 206 surgical technique for, 197–199 technical considerations in, preparation, procedure, postprocedure, 207–210 transcondylar fossa approach (TFA) in, 197 in trigeminal neuralgia (TN), 195–196, 198 tumors of the CPA and, 197 utility of NIOM in, 206 minimal alveolar concentration (MAC). See mean/minimum alveolar concentration monitoring equipment, 9–12, 10, 12, 15, 71, 73. See also NIOM machines; remote monitoring montages, electroencephalographic (EEG) and, 39, 39, 40 morphine, 130 motor evoked potentials (MEPs), 9, 10–11, 14, 15, 17, 21, 29, 97, 128 anesthesia and, 37–38, 38t, 59, 60, 129–130, 239 barbiturates in, 61 benzodiazepines in, 61 compound muscle action potentials (CMAPs) in, 125, 129–131, 133, 134, 135, 148, 149 data displays for, 88–89 electroencephalography (EEG) and, 40, 40 etomidate in, 62–63 interpretation of, 108–111 intracranial pressure and, 58 ketamine and, 62 lumbosacral surgery and, 139 muscle potentials in, 106–107, 107, 108 nerve action potentials (NAPs) in, 125, 126 neuromuscular blockade and, 130 opioids in, 61–62 paralytics in, 63, 64 preparation, procedure, postprocedure for, 111–115, 134–136, 163–166, 251–257 propofol in, 61 radiculopathy in, 110 right-body, during scoliosis correction, 36 spinal cord compromise in, 110–111, 111 spinal cord stimulation in, recording from muscles/nerves, 125 spinal cord surgery and, 117, 124–137, 128 stimulators and, 75–77 systemic factors affecting, 109 technical factors affecting, 109–110 thoracic aortic surgery and, 239, 242–251, 244t, 247, 248 transcranial (tceMEP), 33–38, 77 anesthesia and, 37–38, 38t complications of, 37 contraindications of, 37 degraded signals in, 35, 35 indications for, 37 interpretation of, 35–37 recording of, 34 signal optimization in, 34 stimulation for, 34 techniques for, 34

Husain 18 Index

1/17/08

12:09 PM

Page 311

Ind e x transcranial electric stimulation (TES) and, 106–107, 107, 108, 125, 125, 129, 133 tumors of the CPA and, 218–220 vertebral column surgery and, 96–115 warning criteria for, 108–111, 130–133, 132, 133 motor nerve conduction studies (triggered EMG), 21–29. See also electromyography motor unit action potentials (MUAPs) peripheral nerve surgery and, 186–187 selective dorsal rhizotomy (SDR) and, 171 motor unit potential (MUP), 21, 23, 25–26 multichannel EP machines, 73–74 multiple myeloma, 119–121, 119t muscle potentials, SEP, 106–107, 107, 108 muscles commonly monitored by EMG, 21, 23t muscular dystrophy, 97 MVD. See microvascular decompression myotomes, 21 NAPs. See nerve action potentials NCS. See nerve conduction studies needle electrodes, 16, 173, 186, 186 nerve action potentials (NAPs), 125, 126 anesthesia and, 204 biopsy and fascicle selection using, 190 interpretation of, 204–205 knee neuropathy and, 189 microvascular surgery (MVD) and, 204–207 peripheral nerve surgery and, 182–191, 185 preparation, procedure, postprocedure, 207–210 tumors of the CPA and, 219 tumors of the peripheral nerves and, 189 nerve conduction studies (NCS), 74 amplification in, 81 data displays for, 88–89 microvascular surgery (MVD) and, 207 peripheral nerve surgery and, 183–186 stimulators and, 75 tumors of the CPA and, 219 nerve dysfunction vs. EMG activity, 26–27 networked computers, remote monitoring and, 45–48 neural tube defects, 97, 156. See also tethered cord syndrome neuroblastoma, 119–121, 119t neurofibroma, 119–121, 119t neurofibromatosis, 96 neuromuscular blockade, 130 electromyography (EMG) monitoring of, 147–148, 148, 149, 153 neuromuscular disease, 97 neurophysiologic intraoperative monitoring (NIOM), 3, 21–44 neurostimulator programming and analysis (Codes 95970–95979), 69, 70t Nicolet Endeavor stimulator, 106 NIOM machines, 73–92 amplifiers for, 79–83 analog-to-digital conversion (ADC) in, 83–86, 85 bits and vertical resolution in, 85–86

•

3 11

CD or DVD storage in, 88 comparison of, 89, 90–91t data displays in, 88–89 dwell time in, 84–85 filtering and smoothing in, 87–88 filters in, 82–83, 82, 83 horizontal resolution in, 84–85 operational amplifiers in, 79–80 sampling frequency in, 84–85, 85 signal averaging and, 86–87 signal-to-noise ratio (SNR) and, 86, 87 stimulators in, 74–79 storage in, 88 trending in, 88 NIOM. See neurophysiologic intraoperative monitoring nitrous oxide, 58, 58t, 60, 130, 297 nonintracranial EEG (Code 95955), 69 noninverting op amp, 79–80, 79, 80, 81t nurse practitioner (NP), 7 nursing team, 7–8 Nyquist theorem, 85 Ohm’s law, 76 Ojemann stimulator, 298 oligodendroglioma, 119–121, 119t operating room, 3–19 access to patient in, 15, 16–17 aseptic technique and sterilization in, 3–6 attire for, 5, 17 clean-up of, 17 documentation of procedures in, 17–18 equipment of, 8–13 etiquette and protocol of, 16–17 personnel for, 6–8 preparing for surgery in, 13–15 safety in, 15–16 operational amplifiers, 79–80, 79, 80, 81 common-mode rejection ratio in, 80–81 input impedance in, 81, 81 polarity and inverting/noninverting, 79–80, 79, 80, 81t voltage-divider rule in, 81, 81 opioids in, 60, 60t, 61–62, 130 oscilloscope, 73 ossification of the posterior longitudinal ligament (OPLL), 97, 98 osteoblastoma, 119–121, 119t osteochondroma, 119–121, 119t osteoid sarcoma, 119–121, 119t osteosarcoma, 119–121, 119t pancuronium, 130 paragangliomas, 215t. See also tumors of the cerebellopontine angle paralytics, and anesthesia, 63, 64 pathologic EP decrements, 110, 113, 114 patient access, 15, 16–17 pcAnywhere, 52 pedicle holes and screws, EMG evaluation of, 27–28, 28t, 29, 144–147, 144, 145, 146t

Husain 18 Index

312

•

1/17/08

12:09 PM

Page 312

In d e x

perfusionist, 8 periodic lateralized epileptiform discharges (PLEDs), 240 peripheral nerve stimulation, recording from spinal cord, 122, 123 peripheral nerve surgery, 181–193 common peroneal neuropathy at the knee and, 189 compound muscle action potential (CMAPs) in, 182, 187, 188 compound muscle action potential (CMAPs) in, 183–186, 190–191 electromyography (EMG), needle electromyography in, 186–187, 186 fascicle selection for nerve biopsy and, 190 localizing lesions for, 182 magnetic resonance imaging (MRI) in, 190 median neuropathy at the wrist (carpal tunnel syndrome) and, 189 motor unit action potentials (MUAPs) in, 186–187 nerve action potentials (NAPs) in, 182–191, 185 nerve conduction studies in, 183–186 stimulator placement in, mono- and bipolar, 183–184, 184 technical considerations in, preparation, procedure, postprocedure, 190–191 technical problems encountered in, 187–188 tumors of the peripheral nerves and, 189 ulnar neuropathy at elbow (UNE) and, 188 peripheral neuropathies, 97 peripheral potentials, SEP, 105 personal firewalls, remote monitoring and, 51 personal protective equipment, 16 personnel of the operating room, 6–8 physician assistant (PA), 7 piezoelectric transducers, stimulators and, 78–79 pinging a network connection, 46, 47 plasmacytoma, 119–121, 119t plexus disturbances, 110 plotters, 73 polarity of operational amplifiers, 79–80, 79, 80, 81t polio, 97 port numbers, in remote monitoring computers, 47 portable NIOM machines, 74 positron emission tomography (PET), 296 posterior decompression, 99 posterior spinal fusion (SPF), 98 postoperative documentation, 18 postsurgical testing, 17 potassium/sodium/calcium balance, anesthesia and, 58 potential fade, 109, 109 preoperative documentation, 17 preoperative studies, 13–14 preparing for surgery, 13–15 processing methods, electroencephalography (EEG) and, 40–41

propofol, 60, 60t, 61, 130, 239, 297 protocol in the OR, 16–17 proximal nerve or plexus disturbances, 110 quantitative EEG (QEEG), 271–280, 276 radiation sterilization, 4–5 radiculopathy, 110 free-running EMG and, 142–143, 143t, 143 radiology technician, 8 raw EEG (REEG), 271–280, 274 recording electrodes and parameters, EMG, 150, 150t, 151 reflexive activity, on EMG, 25 regional anesthesia. See anesthesia registered EEG technologist (R EEG T), 70, 71 registered evoked potential technologist (R EP T), 70, 71 registered nurse (RN), 8 Relative Value Units (RVU), in billing/coding, 67 remifentanil, 130, 297 remote application servers, 49–50, 50 remote monitoring, 15, 45–54 client-server systems in, 51–52 computer systems for, 45–48 connecting to the network for, 47, 48–50 data packet used in, 47–48, 48 data transmission in, 50–54 default gateways for computers used in, 47 display software used in, vendor-specific, 50–51 Ethernet connector for, 47, 47 filtering of data in, 47–48 firewalls in network computers used in, 47, 48, 50, 51, 53, 53 fully qualified names of computers in, 51 HIPAA policies and, 53, 88 image capture in, 53–54 in-hospital vs. remote display systems in, 51 Internet and, 45–48 IP (Internet Protocol) address of computers in, 45, 48, 51 keep-alive signals and, 52–53, 53 pcAnywhere and, 52 personal firewalls in, 51 pining a network connection for, 46, 47 port numbers of computers used in, 47–48 remote application servers used in, 49–50, 50 reverse-connection software for, 52–53, 53 screen capture software for, 51–52 security of signal in, 52, 53, 88 servers for, 53–54 virtual network computing (VNC) in, 52 virtual private networks (VPNs) for, 48–49, 49, 51, 52 Web servers for, 53–54 Windows-based systems for, 45–46, 46, 47 repetitive microtrauma of nerve, EMG activity, 26–27, 26 resident, anesthesia, 7 resident, surgical, 6

Husain 18 Index

1/17/08

12:09 PM

Page 313

Ind e x resolution, horizontal, 84–85 resolution, vertical, 85–86 Resource Based Relative Value System (RBRVS), in billing/coding, 67 responsibility of staff, 70–71, 70 reverse-connection software, in remote monitoring, 52–53, 53 reversible ischemic neurologic deficit (RIND), 266 rhizotomy. See selective dorsal rhizotomy roll-off, in filtration, 82–83 room attendant, 8 safety in the OR, 15–16 sampling frequency, 84–85, 85 schwannoma, 119–121, 119t, 215t. See also tumors of the cerebellopontine angle scoliosis, 96–97. See also vertebral column surgery Cobb angle and, 96–97 surgical treatment of, 98 vertebral column surgery and, 96–97 Scoliosis Research Society, 33 screen capture software, for remote monitoring, 51–52, 51 screws, EMG evaluation of, 27–28, 28t, 29, 144–147, 144, 145, 146t scrub nurse, 8 security of data, 52, 53, 88 sedation. See anesthesia selective dorsal rhizotomy (SDR), 169–180 anesthesia and, 172 anesthetic/nerve root verification in, 173 dorsal nerve rootlet selection in, 174–176, 175, 176 dorsal nerve rootlet sparing in, 176–177 electromyography (EMG) and, 171–173 exposure/anatomic identification in, 172 indications for, 170 motor unit action potentials (MUAPs) and, 171 muscle monitoring in, 171, 171t NIOM techniques for, 170 preoperative preparation for, 171–172, 173 somatosensory evoked potentials (SEPs) and, 171 standard montage used for, 173–174, 174t stimulation and recording parameters in, 172–173 subdermal needle electrodes used in, 173 technical considerations in, preparation, procedure, postprocedure, 177–180 threshold ratios (TR) and, 173–174, 175t sensitivity of NIOM to anesthesia, 64–65 SEPs. See somatosensory evoked potentials servers, remote application/remote monitoring, 49–50, 50, 53–54 setup for NIOM, 14–15 sevoflurane in, 58, 58t, 59, 297 shape of MUPs, 26 shock hazards, 15–16 signal averagers, 73, 86–87 signal optimization, MEP, transcranial and, 34

•

3 13

signal rejection, 74 signal-to-noise ratio (SNR), 86, 87, 99–101, 101 smoothing of digital signals, 87–88 somatosensory evoked potentials (SEPs), 9, 15, 21, 29–33, 32, 127, 127t, 132, 133 anesthesia and, 33, 57, 59, 60, 64, 128–130, 239–240, 244, 272–273, 278–279 barbiturates in, 61 blood flow and, 57 carotid endarterectomy (CEA) and, 270–278, 277 cortical potentials in, 100–104 decussation and, 100–104, 101 deep hypothermic circulatory arrest (DHCA) and, 241 distal nerve disturbances in, 110 electroencephalography (EEG) and, 41 epidural spinal cord stimulation in, 124 epilepsy/epilepsy surgery and, 299 feedback rapidity in, 100 filtration in, 82 functional brain mapping and, 296–297 generators of, 30–31, 30t hematologic status and, 57 importance of, 29–30 indications for, 33 interpretation of, 31–33, 108–111 intracranial pressure and, 58 ketamine and, 62 localization of, 30–31, 31t lower limb derivations for, 101–104, 102, 103, 104t lumbosacral surgery and, 139, 140 microvascular surgery (MVD) and, 199, 206 muscle potentials in, 106–107, 107, 108 neuromuscular blockade and, 130, 147–148, 148, 149 paralytics in, 63, 64 peripheral nerve stimulation in, recording from spinal cord, 122, 123 peripheral potentials in, 105 preparation, procedure, postprocedure for, 111–115, 134–136, 163–166, 207–210, 221–226, 251–257, 278–280 propofol in, 61 proximal nerve or plexus disturbances in, 110 radiculopathy in, 110 sampling frequency for, 85 selective dorsal rhizotomy (SDR) and, 171 signal-to-noise ratio (SNR) in, 99–101, 101 “significant” degradation in, 31, 32, 33 spinal cord compromise in, 110–111, 111 spinal cord stimulation in epidural, 124 recording from scalp, 123–124, 124 recording from spinal cord, 122–123 spinal cord surgery and, 117, 122–137, 127, 127t, 132, 133 spinal potentials in, 105 stimulators and, 75 subcortical potentials in, 104–105

Husain 18 Index

314

•

1/17/08

12:09 PM

Page 314

In d e x

somatosensory evoked potentials (SEPs) (continued) systemic factors affecting, 109 technical factors affecting, 109–110 tethered cord syndrome (TCS) and, 160–162 thoracic aortic surgery and, 239–251, 239, 244t, 248, 249 thoracic decompression and fusion shown on, 32, 32 total intravenous anesthesia (TIVA) in, 128, 130 transcranial electric stimulation (TES) and, 106–107, 107, 108 trigeminal nerve monitoring using, 220 tumors of the CPA and, 217–220, 218, 221 upper limb derivations for, 100–101, 101 ventilatory status and, 57 vertebral column surgery and, 96–115 warning criteria for, 108–111, 130–133, 132, 133 sound and the decibel scale, 77–79 sound pressure level (SPL), 78 sound reference and decibel scale, 78 spasticity, selective dorsal rhizotomy (SDR) for, 169–170. See also selective dorsal rhizotomy sphincter function monitoring, 161–162, 161 spinal cord afferent and efferent pathways of, 118–119 anatomy of, 96, 117–119 arteries of, 119 blood supply of, 96 compromise of, 110–111 fiber tracts of, 118–119 spinal cord stimulation epidural, 124 recording from muscles/nerves, 125 recording from scalp, 123–124, 124 recording from spinal cord, 122–123 spinal cord surgery, 117–137 anesthesia and, 128, 129–130 arteries of spinal cord in, 119 carbon dioxide (CO2) laser in, 122 Cavitron ultrasound surgical aspirator (CUSA) in, 122, 135 cerebrospinal fluid (CSF) tests for, 121 clinical presentation in, 121 compound muscle action potentials (CMAPs) in, 125, 129–131, 133, 134, 135 CT and MRI in, 121 diagnostic evaluation for, 121 electromyography (EMG) in, 136 fiber tracts of spinal cord in, 118–119 MEPs in, 117, 124–137, 128 nerve action potentials (NAPs) in, 125, 126 paradigms common to, 125–126 peripheral nerve stimulation in, recording from spinal cord, 122, 123 somatosensory evoked potentials (SEPs) in, 117, 122–137, 127, 127t, 132, 133 spinal cord anatomy in, 117–119

spinal cord stimulation in epidural, 124 recording from muscles/nerves, 125 recording from scalp, 123–124, 124 recording from spinal cord, 122–123 surgical procedure for, 121–122 technical considerations in, preparation, procedure, postprocedure, 134–136, 163–166 total intravenous anesthesia (TIVA) in, 128, 130 transcranial electrical stimulation (TES) in, 125, 125, 129, 133 tumors of spine and, 119–122, 119t utility of NIOM in, 133–134 warning criteria in, 130–133, 132, 133 spinal dysraphism. See tethered cord syndrome spinal potentials, SEP, 105 spinothalamic tract, 118 split-cord malformation, 157. See also tethered cord syndrome spondylosis, cervical, 97, 98 staffing, training and responsibilities of, 70–71 sterile field, 5–6 sterilization methods, 3–6, 17 stimulating electrodes and parameters, EMG, 151, 152, 152 stimulators, 73, 74–79 for auditory stimulation, 77–79 brainstem auditory evoked potentials (BAEPs) and, 78, 79 piezoelectric transducers for, 78–79 sound and decibel scale in, 77–79 sound pressure level (SPL) and, 78 sound reference and decibel scale in, 78 types of, 78–79 bipolar, 183–184, 184 brainstem auditory evoked potentials (BAEPs) and, 78, 79 constant voltage vs. constant current in, 76–77, 76 duration of stimulus and, 75–76, 75 electrical, 75 evoked potentials (EPs) and, 78 intensity of stimulus and, 75, 75 monopolar, 183–184 transcranial MEPs (tceMEP) and, 77 storage for NIOM data, 88 stroke, 263, 266–267, 266. See also transient ischemic attack subcortical potentials, SEP, 104–105 subdermal needle electrodes, 173 sufentanil, 130 supplies for NIOM, 14, 14t surgical documentation, 18 surgical events and the EMG recording, 24–25, 26 surgical resident, 6 surgical tables, 12 surgical team, 6 syringomyelia, 97 systemic factors affecting EPs, 109

Husain 18 Index

1/17/08

12:09 PM

Page 315

Ind e x tables, surgical, 12 tceMEP. See motor evoked potential, transcranial technical factors affecting EP, 109–110 technologists, training and responsibilities of, 70–71 temperature and NIOM, 56 temporal lobe epilepsy (TLE), 284, 285, 291 tethered cord syndrome (TCS), 97, 155–167 anatomy of, 155–156 compound muscle action potentials (CMAPs) in, 158–159, 159 diagnosis of, 157–158 electroencephalography (EEG) in, 158 electromyography (EMG) motor system monitoring during, 158–163, 160 future directions in treatment of, 166–167 paradigm for, 163 pathophysiology of, 156–157, 157t sensory system (SEP) monitoring during, 160–162 sphincter function monitoring during, 161–162, 161 surgical considerations for, 158 symptoms of, 157–158, 157 utility of NIOM in, 162 thermal sterilization, 3–4 thoracic aortic surgery, 227–259 anatomy and, 227–229 anesthesia and, 239–240, 243–244, 246 aneurysmal disease and, 229–231, 229, 230 aortic diseases amenable to, 229–232 aortic dissection and, 231–232, 231 in ascending aortic and aortic arch, 232–233, 232, 233, 238–241, 249–256 bilateral independent lateralized epileptiform discharges (BiPLEDs) in, 240 cardiopulmonary bypass (CPB) in, 232–234, 234, 237–238 cerebrospinal fluid (CSF) pressures in, 242 deep hypothermic circulatory arrest (DHCA) in, 232–250, 234, 240, 241, 241t in descending thoracic and thoracoabdominal aorta, 233–252, 234, 235, 254–257 ECG in, 251 electroencephalography (EEG) in, 239–240, 249–250 endovascular stent grafts (EVSG) in, 235–237, 236, 237, 246–249, 251 endovascular thoracic aortic repair in, 235–237, 236, 237, 246–249, 251 future directions in, 257 generalized periodic epileptiform discharges (GPEDs) and, 240, 240 mean arterial pressure (MAP) and, 243 MEPs in, 239, 242–251, 244t 247, 248 NIOM paradigms for, 238–249 periodic lateralized epileptiform discharges (PLEDs), 240 somatosensory evoked potentials (SEPs) in, 32, 239–251, 244t, 248, 249 technical considerations in, preparation, procedure, postprocedure, 251–257

•

3 15

total intravenous anesthesia (TIVA) and, 243–244, 246 utility of NIOM in, 249–251 thoracic decompression and fusion, somatosensory evoked potentials (SEPs) and, 32 threshold ratios (TR), EMG, 173–174, 175t total intravenous anesthesia (TIVA), 7, 15, 60, 61, 62, 64, 128, 130, 243–246 train-of-four (TOF) blockade, 63, 147–148, 148, 149 training, 70–71 trains, EMG, 23, 24 transcondylar fossa approach (TFA), 197 transcranial Doppler (TCD), carotid endarterectomy (CEA) and, 270 transcranial electric stimulation (TES), 106–107, 107, 108, 109–112, 125, 125, 129, 133 transcranial electrical motor evoked potentials, 33–38 transesophageal echocardiography (TEE), 10–11, 10 transient ischemic attack (TIA)/stroke, 263, 266–267, 266 trauma, spinal, 97 trending, 88 trigeminal nerve monitoring, 219–220 trigeminal neuralgia (TN), 195–196, 198. See also microvascular decompression tumors of the cerebellopontine angle (CPA), 197, 213–226 anatomy and, 213–215 brainstem auditory evoked potentials (BAEPs) and, 214, 216–220, 216, 221, 222 compound muscle action potentials (CMAPs) and, 219, 224 electroencephalography (EEG) in, 222 electromyography (EMG) and, 217–220, 221, 225 evoked potentials (EPs) and, 219 facial nerve monitoring in, 219 magnetic resonance imaging (MRI) and, 215 MEPs and, 218–220 nerve action potentials (NAPs) and, 219 nerve conduction studies (NCS) and, 219 NIOM techniques in, 217–220, 220t pathology of, 215, 215t somatosensory evoked potentials (SEPs) and, 217–220, 218, 221 surgical procedures for, 215–216 technical considerations in, preparation, procedure, postprocedure, 221–226 timing of neural injury in surgeries of, 216–217 trigeminal nerve monitoring in, 219–220 tumors of the peripheral nerves, 189 tumors of the spine, 97, 117, 119–122, 119t Turner’s syndrome, 97 ulnar neuropathy at elbow (UNE), 188 ultrasound, 10–11, 10 Cavitron aspirator (CUSA), 122, 135

Husain 18 Index

316

•

1/17/08

12:09 PM

Page 316

In d e x

ultraviolet (UV) protection, 16 upper limb derivations (SEP), 100–101, 101 urethral sphincter function monitoring, 161–162, 161 vagus nerve, 262, 264 vascular surgeries, 276–277 vascular tumors, 215t. See also tumors of the cerebellopontine angle vecuronium, 130 ventilation during anesthesia, 57 ventral posterolateral (VPL) nucleus, 118 vertebral column surgery, 95–116 anatomy of vertebral column in, 95 anesthesia for, 112–113 anterior cervical discotomy and fusion as, 98 anterior spinal release as, 98 baseline readings for, 113 clinical features of disorders in, 97 cortical potentials for, 100–104 decussation and, 100–104, 101 distal nerve disturbances in, 110 electroencephalography (EEG) in, 101 evoked potentials (EP) in, 95 feedback rapidity in, 100 interpretation and warning criteria in, 108–111, 109 lower limb derivations (SEP) for, 101–104, 102, 103, 104t monitoring techniques in, 99 motor evoked potentials (MEPs) in, 96–115, muscle potentials in, 106–107, 107, 108 pathologic EP decrements in, 110, 113, 114 peripheral potentials in, 105 posterior decompression as, 99 posterior spinal fusion (SPF) as, 98 preparation, procedure, postprocedure for, 111–115 proximal nerve or plexus disturbances in, 110

radiculopathy in, 110 scoliosis in, 96–98 signal-to-noise ratio (SNR), 99–101, 101 somatosensory evoked potentials (SEPs) in, 96–115 spinal cord anatomy in, 96 spinal cord blood supply in, 96 spinal cord compromise in, 110–111, 111 spinal potentials in, 105 subcortical potentials in, 104–105 surgical feedback tracing, 113 systemic factors affecting, 109 technical factors affecting, 109–110 transcranial electric stimulation (TES) and, 106–107, 107, 108 upper limb derivations (SEP) for, 100–101, 101 vertical resolution, 85–86 video-EEG, 287, 289 Viking NIOM stimulator, 106 virtual network computing (VNC), 52 virtual private networks (VPNs), for remote monitoring, 48–49, 49, 51, 52 voltage-divider rule, amplification, 81, 81 Wada test, 296 Web servers, remote monitoring and, 53–54 widespread anteriorly maximum rhythmic activity (WAR) in, 272 widespread persistent slow waves (WPS), 272, 273 Windows-based remote monitoring systems, 45–46, 46, 47 wrist, median neuropathy (carpal tunnel syndrome) at, 188, 189 X-ray machines, 11, 16 Z-meters, 177